Colonial Development and Population in Taiwan 9781400877010

Citation preview

Colonial Development and Population in Taiwan

UNDER THE EDITORIAL SPONSORSHIP OF THE OFFICE OF POPULATION RESEARCH PRINCETON UNIVERSITY

COLONIAL DEVELOPMENT AND POPULATION IN TAIWAN BY GEORGE W. BARCLAY

PRINCETON, NEW JERSEY PRINCETON UNIVERSITY PRESS 1

954

Copyright, 1954, by Princeton University Press London: Geoffrey Cumberlege, Oxford University Press L.C. Card 52-13153

Printed in the United States of America by Princeton University Press, Princeton, New Jersey

Preface To THOSE outside the field of population studies a book on Taiwan's demographic development must seem an example of academic es­ capism. When even the general reader is becoming aware of the diffi­ cult population problems facing Asia's mainland, why should the Office of Population Research devote personnel and resources to the careful study of a few million people on a rather small island in the Far East? The question is reasonable, but the answers are compelling. Tai­ wan is the only place where it is possible to study the processes of change of a Chinese population over a substantial period of time on the basis of excellent data. Any information that can be gained about the growth characteristics of one of the world's largest ethnic groups is well worth obtaining. Under Japanese rule an unusually accurate statistical record was maintained in Taiwan from 1905 to 1943, providing an unbroken series of data that is longer, and in some respects richer, than that for Japan proper. Indeed, the Taiwanese material provides much more than a unique demographic record of Chinese people. It is, if not the only case, then certainly one of the very few cases in the world in which the trends in fertility and mortality can be traced with precision from the initial stages of modernization. In virtually every other case the statistical record begins long after death rates have begun to decline. In Taiwan, the vital record can be traced over almost the whole of the period during which the island was being transformed from one of the least healthy to one of the most healthy places in the Far East. At a time when much attention is being given to problems of economic development and to discussion of its interrelations with population growth, the Japanese experience in Taiwan is especially appropriate. The Japanese wanted a source of food and a market for the home industries, and their program for the island was di­ rected toward achieving these ends. Here, then, is an opportunity to study the impact of an aggressive program of agricultural develop­ ment on population growth, and the limitations imposed by the nature of such a program on subsequent economic development when the objectives changed, as they did under the strain of war preparation. ν

PREFACE

Mr. Barclay's book is not intended to be another regional study. It is designed to trace the demographic record in relation to the record of social and economic change. It is the fourth volume spon­ sored by the Office which deals with problems of non-industrial regions. The others are: Kingsley Davis, The Population of India and Pakistan (Princeton, 1951); Τ. E. Smith, Population Growth in Malaya (Chatham House, 1952); and Paul K. Hatt, Backgrounds of Human Fertility in Puerto Rico (Princeton, 1952). A fifth volume by Harvey Leibenstein, Theory of Economic-Demographic Develop­ ment, is being issued by Princeton University Press at about the same time as the present book. Mr. Barclay's work was made possible by the generous support which the Office receives from the Milbank Memorial Fund and The Rockefeller Foundation, neither of which, however, has the slightest responsibility for the conclusions reached or, indeed, for the specific topic selected for study. For this opportunity to work in complete freedom, my colleagues and I wish to express our gratitude. Frank W. Notestein Director Office of Population Research November 1953

Introduction TAIWAN, or "Formosa," as the island is still widely known to West­

erners, has been the scene of international contention as long as it has had a recorded history. From its discovery by curious Chinese and Japanese before 1600 down to the present time, the island has repeatedly been visited, occupied and fought over. No one can claim that Taiwan has failed to receive its full share of attention by the outside world. The people of the island would be the last to make this complaint, for, while opinions may differ as to whether the in­ trusions have been beneficial to the population, it is certain that they have not been welcome. But for people who have long felt the disadvantages of being small and weak with large and powerful neighbors, the inhabitants of Tai­ wan have in some respects fared remarkably well. They avoided, through hardly more than chance, the domination by Western nations that has called forth great bitterness and aroused quixotic nation­ alism in many parts of Asia. They fell under the control of Japan at a time when that country was willing to invest strong efforts in de­ veloping Taiwan's productive resources. So far they have emerged from the war that destroyed the Japanese Empire without suffering the mass bloodshed or wide devastation that followed the regular military operations elsewhere in the Far East. And, ironically, during the fifty years of "exploitation" and "oppression" by Japan, they en­ joyed more peace and prosperity than their former countrymen on the mainland of China. As the aftermath of these events Taiwan has acquired a status that is uneasy and insecure, poised again as a prize desired by rival pow­ ers. It is not the purpose of this book to trace the sequence of steps leading up to this situation, or to speculate about the future inter­ national status of the island. Because prominence has not come to them of their own seeking, the Taiwanese have usually played a pas­ sive part in shaping their position in world affairs. Knowledge of the island's external relations therefore does not offer much in under­ standing the social order of its people, and information about the internal structure of the society fails to contribute anything of great value in appreciating its present international position. For similar reasons, this book will not be concerned with the pros­ pects for the island under the present Chinese Nationalist govern-

INTRODUCTION

ment, since they are likewise closely allied to international develop­ ments that have not yet taken place.1 This present era of Taiwan's history has just begun; discussion of its future in these terms neces­ sarily rests on conjecture and may have little to do with the facts that are known about the population. The half-century of colonial rule, on the other hand, is over, and most of the facts are in. Hence this is the phase that can be investigated to the best purpose, for it can serve as an instructive example of its type in other places. It is a well-documented case. There is no lack of thoughtful re­ search on the island during this period. While under Japanese rule Taiwan probably had the distinction of being the most thoroughly inventoried colonial area in the world. Huge compilations of sta­ tistics and numerous special surveys were made from year to year. The economy, the terrain, the aboriginal tribes, the mineral wealth, the agricultural output, the industrial production and the foreign trade have all been studied and restudied until there is little to be added to this knowledge unless new evidence is uncovered that is not now available.2 Strangely enough, with all this interest in finding out about the island, there were very few attempts to study the social institutions and cultural patterns of the people, and the attention devoted to these matters has been scanty and perfunctory. Yet the numerical data relating to the population are the best of all, since their collec­ tion was most directly linked to the very effective system of control imposed by the Japanese. The largest task required of this investiga­ tion has been to test and evaluate these unknown source materials, and to introduce them into the Western literature. They comprise one of the most ample and creditable records for a population of this size that has ever been at the disposal of demographers, and the only systematic set of data for a portion of the Chinese population. This is not to say that the subject of Taiwan's population has been entirely neglected. Nearly every published work about Taiwan, espe­ cially since 1945, has included a section on the people of the island. 1 There is growing literature on Taiwan dealing with its political status. See, for example, Riggs, F. W. (1952), Ravenholt, A. (1952), Bate, Η. M. (1952). 2 Some of this material has already appeared in the form of special studies. See Yanaihara, Tadao (1929), Grajdanzev, A. J. (1942), Ch'en Cheng-hsiang (1950), Taiwan Keizai Nempo (1941). For a study of the characteristics of Taiwan's population, see Ch'en Cheng-hsiang and Tuan Chi-hsien (1951). The most useful single compilation of statistics is the Statistical Summary of Taiwan for the Past 51 Years (Taipei, 1946), which will be referred to hereafter simply as Statistical Summary.

INTRODUCTION

Such discussion is often confined, however, to establishing how many of them there are, and frequently begins and ends with the premise that there are "too many." This issue lies afield from the topics that we propose to consider—a population could not attain its present size, flourish or decline, or even remain in existence were it not for the complex societal arrangements that have made these things pos­ sible. Population thus could not become a "problem" by itself, but is a fruitful way of viewing problems in these other spheres. These, along with trends in the numbers of people, are the subject matter of this study, and the succeeding pages are an endeavor to discover what Taiwan's demographic data can be made to reveal about the nature of the society. The following plan of procedure has been adopted. We shall first describe how Taiwan came to be settled, and take note of the growth of its population, for these are features which mark it off at the out­ set as a special type of case and furnish clues as to what else one should expect to find. We shall then turn to examine the aims and policies of Japanese colonial development and their consequences for the population. It is not proposed to dwell on what is commonly treated as the "economic" side of this colonial program, such as the spread of the Japanese banking system, the corporate structure of their industry, the expansion of output, construction of factories, or contribution to the "welfare" of the population. A catalog of these direct achievements gives but a very superficial idea of their most interesting aspect—the form in which these innovations were com­ municated to the Taiwanese. It will appear that participation in this program on the part of Taiwanese was impeded by barriers at nearly every point, and that they felt its effect on their livelihood and their customs only by in­ direction. Some of the reasons for the cleavage between these two groups are to be found in the restrictive measures of the Japanese, and some in the stability of Taiwanese social institutions. Evidence of this stability will be shown in the economic activities of the peo­ ple, their immobility, the persistence of certain mortality patterns, and the organization of marriage and family affairs. These various aspects of Taiwanese life will also be drawn upon to see how and where minor social changes tended to be fitted into this agrarian sys­ tem. Finally, we shall assess one overshadowing result of this social stability, the steady high fertility of the Taiwanese. Throughout, of course, much of the interest in the Taiwanese as

INTRODUCTION a people stems from the fact that almost all of these are of Chinese origin, and that their behavior under Japanese colonial adminstration may be said to represent the reaction of Chinese cultural norms to selective innovations which fail to attain the dimensions of sweep­ ing societal changes. But this is a pattern of change that has not been uncommon elsewhere in the world. Additional significance can be attached to those findings in Taiwan which have counterparts in other agrarian countries, and the following chapters have been or­ ganized with considerable care to emphasize these matters of more general interest. G.W.B. September 1952

ACKNOWLEDGMENTS FOR a long and patient investment in the work involved in this study, full recognition and thanks are to be given to the Office of Population Research, Princeton University. Invaluable counsel was received from individual members of this office, in particular from Irene B. Taeuber, Frank W. Notestein, and Wilbert E. Moore, which should be recorded here for want of a more adequate form of ac­ knowledgment. Special thanks are due to Daphne Notestein, for dis­ tinctive execution of the maps and diagrams, and to Erna Harm, foremost among several, for her care in the statistical computations on which the study is based. Finally, for contributions of assistance and encouragement at every stage, a debt of a broader kind is owed to Audrey N. Barclay. Of course, responsibility for all statements and conclusions in the book rests with the author.

χ

Contents Preface Introduction Chapter I. The Growth of Population

ν vii 3

Chapter II. "Economic Development" in Taiwan

18

Chapter III. Economic Activities and Social Change

43

Chapter IV. The Disposition of Manpower

73

Chapter V. Migration and the Growth of Cities

102

Chapter VI. Public Health and the Risks of Death

133

Chapter VII. The Setting for Family Life

173

Chapter VIII. Patterns of Marriage and Divorce

210

Chapter IX. The Fertility of the Taiwanese

238

Chapter X. Conclusion

255

Bibliography

263

Index

272

Tables ι. Taiwanese Born Outside of Taiwan, 1920 and 1930 2. Growth of Population in Taiwan 3. Ethnic Composition of the Population of Taiwan, Censuses of 1905 and 1935 4. Value of Exports from Taiwan, by Place of Destination, 1899 to 1942 5. Estimated Per Capita Value of Foreign Trade of Far Eastern Countries, Circa 1939 6. Improvement in the Yields of Rice in Taiwan, 1916 to 1945 7. Gross Recorded Value of all Principal Types of Production, 1915-1942 8. Gross Value of Production Recorded for "Industry" in Taiwan, by Line of Product, 1921-1942 9. Spread of Double-Cropping in Irrigated Fields, 1917 to 1945 10. Labor Requirements for Rice Cultivation in Taiwan 11. Occupational Composition, Total Population of Taiwan 12. Occupational Composition of Working Males in Taiwan. Total, All Groups 13. Relative Growth of Employment in Different Lines. Total Population, All Groups 14. Occupational Composition of Taiwanese 15. Occupational Composition of Japanese in Taiwan 16. Total Males, and Occupied Males Distributed by Occupation, at Census Dates, 1905 to 1930. Total, All Groups, Taiwanese, Japanese, and Others 17. The Labor Force in the Total Population, Taiwan and the United States 18. Age Composition of Taiwan and Selected Other Countries, Circa 1930. Both Sexes 19. Total Population of Taiwan, and Density, 1930 20. Total Population by Size of Civil Division, 1920 and 1935 21. Net Internal Migration of Taiwanese, 1925-1930 22. Inter-Prefectural Migrants in Taiwan, 1930 23. Total Surviving Migrants between Prefectures, and the Portion That Went to Cities 24. Taiwanese and Japanese in Nine Major Cities, at Successive Census Dates, 1915-1940

Page 11 13 16 21 33 36 38 38 53 54 58 60 61 66 67

71 82 99 104 105 111 113 114 116

CONTENTS

Page 25. Proportions of Taiwan's Population of Each Ethnic Group, Living Inside the Nine Major Cities in Taiwan, and in Taipei City, 1920 and 1940 26. Ethnic Composition of Major Cities in Taiwan, 1920, 1930 and 1940 27. Residential Segregation in Taipei City, 1925 and 1935 28. Residential Segregation in Six Cities of Taiwan, 1935 29. Age and Sex Composition of Taiwanese, Inside and Outside of Nine Major Cities, 1925 and 1935 30. Differences in Age and Sex Composition of Taiwanese by City, 1935 31. Inter-City Differences in Occupational Composition 32. Occupational Composition in Cities of Taiwan by Ethnic Group, 1930 33. Occupational Distribution of Taiwanese Males Reporting an Occupation, Inside and Outside Major Cities, 1905, 1920, 1930 34. TestforAdequacyofDeathRegistrationofTaiwanese 35. Test for Adequacy of Birth Registration of Taiwanese 36. Crude Death Rates of Taiwanese, 1906-1943 37. Mean Life Expectancy of Taiwanese at Different Ages 38. Taiwanese Survivors through Early Childhood 39. Infant Mortality of Taiwanese. Deaths per 1,000 Live Births in Each Year 40. Mean Life Expectancy of Taiwanese at Different Ages, by Prefecture, 1924-1925 41. Differences in Mortality by Prefecture, in Selected Years 42. Relative Variation of Mortality among Prefectures in Selected Years 43. Size of Households in Taiwan, Successive Censuses, 1905^94 0

44. Per Cent of Taiwan's Population Living in "Ordinary" Households, 1905-1930 45. Size of Households, by Type and by Nationality of Head 46. Composition of Taiwanese Households, 1920 and 1930 47. Extent of Family Relationship in Taiwanese House­ holds, 1920 48. Family Relatedness and Sex Composition of House­ hold Residents, by Size of Household, 1920 49. Taiwanese Living Alone, 1920 50. Composition of Taiwanese Households—Total, Head Occupied and Head Not Occupied, 1920

118 119 123 124 126 127 128 129

130 143 144 146 154 158 161 168 169 170 1 Ib

177 178 182 186 188 189 191

CONTENTS

Page 51. Composition of Taiwanese Households, by Occupation of Head, 1930 52. "Outsiders" in Taiwanese Households, 1920 53. Proportions of Females in Taiwanese Households, 1920 54. Employment and Dependency in Taiwanese House­ holds, 1920 55. Average Size of Household and Extent of Male Em­ ployment in Families Classified by "Occupational Status," 1920 56. Changes in the Extent of Male Employment in Taiwanese Families, 1920 to 1930 57. Concentration of Employment within Households, 1930 58. Concentration of Employment within Households: Variation between Prefectures of Taiwan, 1930 59. Proportions of Taiwanese Ever-Married, by Age and Sex, 1905-1935 60. Sex Balance of Taiwanese in Marriageable Ages, 1905^940 61. Excess of Married Women in Taiwan, 1905-1935 62. Widows and Widowers in Taiwan 63. Divorced Men and Women in Taiwan 64. Marriages Ended in Divorce, 1906-1940 65. Remarriages of Widowed and Divorced Persons, 1920 66. The Shift toward First Marriages in Taiwan, 1906-1943 67. Relative Restriction of Choice in Mates by Men and Women Marrying for the First Time, 1906-1943 68. Relative Advantage of Men and Women in Remarriage, 1906-1943 69. Per Cent of All Registered Marriages between Taiwanese That Were Matrilocal, 1906-1943 70. Per Cent of Taiwanese Women Currently Married, by Age, 1905-1935 71. Per Cent of Taiwanese Women Married, 1920, by Age and Ancestral Place of Origin 72. Per Cent of Taiwanese Women Married, by Age, Inside and Outside Major Cities, 1925 and 1935 73. Variation in Proportions Married, by Prefecture, Inside and Outside Major Cities, 1925 and 1935 74. Crude Birth and Death Rates of Taiwanese, 1906-1943, and Rates of Natural Increase 75. General Fertility Ratios for Taiwanese, 1905 to 1940

192 194 195 198

202 204 206 207 211 212 213 217 219 221 223 225 226 227 229 231 232 233 234 241 243

CONTENTS

76. Gross Reproduction Rates for Taiwanese, 1905 to 1940 77. Gross Reproduction Rates for Selected Countries 78. Gross Reproduction Rates in Taiwan, 1930 and 1935, Prefectures and Cities Appendix to Chapter VI. Taiwanese Life Tables. Selected Years

Page 246 247 251 172

Figures ι. Topography of Taiwan (Map) 2. Distribution of Population in Taiwan, 1935 (Map) 3. Age Composition of the Taiwanese Population, 1905 and 1940 4. Composition of the Gross Value of Agricultural Production, Principal Crops, 1915-1945 5. Area of Cultivated Land in Taiwan, 1910-1944 6. Expansion and Use of Railroad Facilities in Taiwan, !900-1945 7. Expansion of Irrigation and Agricultural Population in Taiwan, 1905-1947 8. Expansion of Electric Power Capacity and Factory Employment in Taiwan, 1910-1945 9. Crop Production and Cultivated Land Per Person Dependent on Agriculture, Taiwan, 1916-1945 10. Increase of Factory Employment and of Population, 1914-1941 11. Trends in Occupations, 1905-1940 12. Occupational Composition of Taiwanese, by Age, 1905-1930 13. Occupational Movement of Taiwanese Males, 19201930, Into and Out of Agriculture 14. Occupational Movement of Taiwanese Males, 19201930, Outside of Agriculture 15. Occupational Movement and "Inheritance" of Occupa­ tions by Taiwanese, 1920-1930. Net Movement between Occupations, and Movement into Occupa­ tions by Pre-Labor Force Males 16. Taiwanese Population in Working and Dependent Ages, 1905, 1940 and Projected to 1950 17. Net Migration of Taiwanese between Civil Divisions, 1925-1930 (Map) 18. Net Migration and Natural Increase by Civil Division, Related to Intercensal Growth of Population, !925-1930 19. Growth of Population in Major Cities of Taiwan, 1920-1940. Taiwanese and Japanese 20. Residential Segregation of Japanese in Taipei City (Map) 21. Crude Birth and Death Rates, and Total Registered Births and Deaths of Taiwanese, 1906-1943

Page 8 9 14 34 36 39 40 42 55 64 69

8

3

89 93

96 100 108

no 117 122 147

CONTENTS

Page Mortality of Taiwanese, Life Table of 1909-1911 23· Life-Table Survivors of 10,000 Live Male Births, by Age. Taiwan and Selected Other Countries 24. Mortality of Taiwanese by Age (qx). Selected Life Tables, Males and Females 25· Age-Specific Death Rates of Adults in Selected Years. Taiwanese Males and Females 26. Sex Selectivity of Mortality in Taiwan and Other Countries 27. Fall of Death Rates of Infants by Month of Age, 22.

1906-1940

Decline of Mortality by Age, 1906-1941. Taiwanese Males 29· Mortality of Taiwanese Males in Each Age, by Prefecture, 1924-1925 30. Size of Households in Taiwan, 1905-1930 31· Marriage Rates of Taiwanese, 1919-1921 and 1934-1936 32· Gross Reproduction Rates of Taiwanese, by Civil Division, 1935 (Map)

150

151 152 155 156

163

28.

165

167

179 215 250

Colonial Development and Population in Taiwan

CHAPTER I

The Growth of Population IN the last century and again in modern times, Taiwan has repeat­ edly become the object of rivalries among some of the most powerful nations of the world. Though its history has been a succession of disruptive incidents resulting from such contacts, by its size alone the island scarcely seems to merit the covetous attention that out­ siders have given to it. It is not large—only 250 miles from north to south, and about 90 miles across at the widest point. The area that it covers is only 14,000 square miles, or slightly less than that of the Netherlands. There is little in the way of precious minerals that would arouse the desire of industrial countries. It offers no great area into which other peoples can expand. It has only its significant location in the path of transportation routes off the coast of China, and a territory potentially rich in agricultural output. While the former of these has had the greater attraction to foreign nations, it is the latter that actually yielded—perhaps exceeded—the full meas­ ure of wealth that was expected. The topography and climate of Taiwan in the past have combined to make the island inhospitable to settlement. Being at the fringe of the tropics, the climate is hot and rainy, with marked changes of season and variations from one section to another. The greater part of its surface is not readily put to human use. Over half of this area consists of rugged mountains, rising steeply from a plain along the western coast and extending to the Pacific Ocean in the east. Heavy rainfall therefore runs off very quickly, creating torrential floods and occasional droughts in between. Large parts of the island are fre­ quently swept by typhoons and shaken by earthquakes. Within the span of recorded history, the difficulties presented by the physical environment never succeeded in discouraging human habitation. There were vague reports of residents there before A.D. 1000. When Taiwan first came officially to the attention of China, during the Ming dynasty, some tribes of primitives were known to be living near the western coast. Since the 16th century people swarmed to Taiwan from the relatively crowded parts of South China, and prospered there as agriculturists. The Japanese, who ruled the island from 1895 to 1945 and exploited its possibilities

THE GROWTH OF POPULATION

more thoroughly, developed it into a highly profitable colonial en­ terprise. There was some solution to each of its principal deterrents to settlement; but the deliberative and organized approach that many of these solutions required was adopted only by the Japanese. Under them production expanded and population increased at a rate that was prodigious, until the very success of these solutions created its own uncertainty—how and when the increase of people would cease. Early Settlement in Taiwan

Piecing together the history of Taiwan's population before 1895 is a difficult task. It cannot be pursued very far, since relevant in­ formation is lacking. There is no reliable record to show even the number of people on the island before the 20th century. The dimen­ sions of population growth, and its nature, must therefore be sur­ mised from the reports of travelers, tradesmen and Western mis­ sionaries. Prior to 1600 Taiwan was sparsely occupied by a few tribes of "savages," of origins that are unknown. The island at this time had attracted very few people from China, though its mountains were actually visible from the mainland, and its coast less than a full day's sail with the tide. Some Chinese and Japanese used it as a base for commerce and piracy against other shipping along the coast. The early tribes of non-Chinese ranged over most of the inhabited area, sometimes repelling visits by curious foreigners, sometimes warring among themselves. These were forebears of the present aborigines of the interior, the "real" Taiwanese. At about this time the Dutch entered regular and quasi-legitimate trade in the Far East, and became interested in Taiwan as a base for this purpose. In 1622 they established a few fortified outposts on the island, commenced trading across the straits, and formed a limited government to minister to what they took to be the civil and re­ ligious needs of the native population. Their rule left nothing of lasting importance in Taiwan except the early chronicles that they faithfully kept of their experiences, which form the safest repository of information about contemporary events. From such sources we are informed of a growing influx of migrants from the mainland. These were a special category of Chinese, the so-called Hakka people from Kwangtung Province. They were an ethnic minority in China, and as such they occasionally were targets

THE GROWTH OF POPULATION

of organized abuse. Following an outbreak of such disturbance in Kwangtung a considerable number went to Taiwan for refuge, where they were not easily persuaded to respect Dutch rule, and hence were not eagerly received. But even while Taiwan was nominally in the hands of a Western power, these migrants represented a substantial Chinese settlement there. The Dutch were not able to remain very long in Taiwan. As an aftermath of a change of dynasties in China, the slow stream of Hakka people became engulfed in an organized movement of a defeated faction from South China. Led by the celebrated personage Cheng Ch'eng-kung, the newcomers drove out all Westerners from Taiwan in 1662, only to succumb after some years to control by the new Manchu regime in China. These events were followed by continuous migration of Chinese. Most of this flow started from Fukien Province, the new base for administration of Taiwan. These new migrants soon outnumbered by far the previous settlers. Thereafter, the Chinese coming into Taiwan gradually spread out over the flat area most suitable to intensive farming. Little by little they displaced aborigines from the lowlands, and pushed them back into the mountains. Steady movement from the mainland raised the population of Taiwan from a few thousand in 1600 to more than two million by 1895. The occupied area was enlarged during this period from a few scattered points along the western coast to the entire level portion of the island.1 More important, in the process of this settlement Tai­ wan was changed into a land that was thoroughly Chinese, imprinted permanently with the characteristics of Chinese culture. The abo­ riginal people of Taiwan became a small minority, and remained apart in the interior where they could better hold off further incur­ sions on their territory.2 Hence the "Taiwanese" that one can speak of at any later date consisted almost entirely of these early migrants from China and their descendants. By the time Taiwan was ceded to Japan, this population had made it culturally an extension of China. From the sphere of politics and economic activities, as we shall see in later chapters, to matters of 1 See Ch'en Cheng-hsiang, Atlas (1950), Fig. ig. 2 They made good use of the advantages of rough terrain in protecting themselves. In spite of the fact that the tribes quarreled among themselves and failed to present a united front against conquest, it later took the Japanese over thirty years to subdue this interior part of the island, in a campaign with modern weapons, airplanes, and re­ portedly even poison gas.

THE GROWTH OF POPULATION

village life and family relationships, the people of Taiwan retained the heritage of tradition and custom that they had brought from the mainland. These cultural institutions did not succeed so well in maintaining order in Taiwan as in China, for there were new situa­ tions to be met and resolved. Chief among these, for example, was the constant arrival of new persons for whom the obligations of a particular locality might not be binding. But once established in Taiwan, there is no evidence that these institutions underwent any major change throughout the 200 years while Manchu authorities ruled the island. Historical accounts of this period are little but repetitions of the same events. They show a sequence of misrule, intrigue by privi­ leged groups or classes, frequent bloody insurrections, and strife with the aborigines. The island gained a continual atmosphere of violence and unrest. By one important circumstance the disorders were pre­ vented from being an invitation to social change—they were not revolutions, but rebellions. They arose from dissatisfaction with par­ ticular persons in the social order—landlords, officials, or members of rival clans or villages—rather than dissatisfaction with the social order itself.3 As a result, the disorders often served in the end to establish tradition more firmly. Taiwan under the Japanese

The most decisive event in the history of Taiwan after being opened to Chinese immigration was its annexation into the Japanese Empire. Japan in the 1890's was newly arrived as a nation to com­ pete in Far Eastern political rivalries, and many of its leaders were eager to find areas for its expansion as a respectable colonial power. They magnified a relatively minor incident in Taiwan, together with other grievances, into a pretext for provoking war with China in 1894.4 It was a brief war of several quick military victories for the Japanese. When peace was negotiated in 1895, they insisted that Tai­ wan be ceded to them, and in the same year they took possession of 3 Rebellions were ultimately settled by reasserting the established order, though often not without the use of force in the form of troops. Many of them, however, were actually organized within its framework from the beginning. Much of the disturbance was ap­ parently created by the type of large clan organization found in South China, and by secret societies that were able to develop rivalries on a large scale between formidable groups. 4Japanese forces had as early as 1874 made an exploratory expedition to Taiwan, under a similar pretext, that formed the basis for expansionist designs on the island and provided valuable information about its terrain.

THE GROWTH OF POPULATION

the island and set out to discover the responsibilities and rewards of colonial administration. Japan embarked on this first colonial venture without having at the outset any very clear idea of what use to make of it. The first few years were taken up in consolidating their control over the popu­ lation and in deciding what to do with the island. As their experi­ ment progressed, the Japanese began to introduce some profound and subtle changes. Their objectives became clarified, and their poli­ cies toward Taiwanese more firm. Gradually they evolved a system of managing affairs in Taiwan that developed into one of the most successful colonial programs in the world. As long as they remained, Taiwan repaid them handsomely for their initial investment of care and expense. Some of the more important aspects of this program will be dis­ cussed in subsequent chapters. Here it is appropriate to establish what were broadly the consequences in terms of population trends, and to discover what sorts of information about Taiwan's people can be gained from the records of this administration. The Japanese rationalized certain parts of Taiwan's agriculture, so that the island produced a substantial surplus of farm commodities each year. They established a strong and efficient government, the first that the island ever had. With a shrewd combination of police force and political guile, they imposed strict public order and penetrated every town and village with a structure of organized control. They constructed excellent facilities for transportation and com­ munication where there had been none before, founded modern business institutions and commercial centers, promoted some indus­ trial processing of the island's products, rebuilt the major cities, and stamped out the principal epidemic diseases. In short, they trans­ formed Taiwan from a "backward" and neglected land into a thriv­ ing region that could regularly export a large share of its agricultural produce. This was a success that would satisfy most of the countries striving for modernization today. All these material accomplishments of the Japanese in Taiwan are well known. With some justifiable pride, they publicized this side of their rule and compiled an imposing set of statistics to docu­ ment it. What we propose here is to examine the human side of these developments and to appraise some of its consequences in the

THE GROWTH OF POPULATION

TAIWAN, ELEVATION

TROPIC OF1 CANCER

METERS O - 499 500-1,999 2000 50 MILES

OFFICE OF POPULATION RESEARCH, PRINCETON UNIVERSITY

Fig. 1. Topography of Taiwan. (Adapted from Taiwan Provincial Government, Department of Communications, A Map of the Main Roads of Taiwan, February, 1950.)

lives of the population. This aspect of Taiwan's history, and of "eco­ nomic development" in general, receives a surprisingly small share of attention as compared with questions of technological advance.

THE GROWTH OF POPULATION

DISTRIBUTION OF POPULATION TAIWAN, 1935 (EACH DOT REPRESENTS APPROXIMATELY 2 , 5 0 0 PERSONS)

'.'AS

TROPIC OP CANCER

SO MILES

ZZ

OFFICE OF POPULATION RESEARCH, PRINCETON UNIVERSITY

Fig. 2. Distribution of Population in Taiwan, 1935. All nationality groups,

The Japanese and Statistics of the Population

The only trustworthy information that we have about the num­ bers and kinds of people in Taiwan at any time is the evidence of the censuses taken by the Japanese. They conceived and established

THE GROWTH OF POPULATION

a system of collecting population data, as an integral part of their wider goals in the island. Factual materials of this order have not existed either before or since their occupation. The complete monument to their enumerative capacities in Tai­ wan is a series of seven censuses, the first one in 1905, the next in 1915, and one every five years thereafter stopping with 1940. Some were more detailed and are therefore more useful than others. All maintained uniformly high standards in the completeness of report­ ing, in the consistency of defining ethnic classes, and in the ac­ curacy of information such as age reporting. In these respects they often ranked in quality well above the more advanced censuses of the Western world. The first two censuses, in 1905 and 1915, were carried out before a genuine enumerative census was ever conducted in Japan Proper5 and probably served as testing-grounds for the initial Japanese census of 1920. One of the first official acts of the Japanese was to make a register of the population under their control. This was undertaken for pur­ poses of regulation, and not to satisfy disinterested curiosity about the inhabitants. To judge by its published results, the early registry was incomplete, especially in its vital statistics. But the rules re­ garding nationality were important for later demographic data, be­ cause they provided the first basis for distinguishing ethnic groups in the population. The authorities began by requiring all persons who acknowledged the legal status of Taiwanese "nationals" to be so documented with the police. The remaining people of Chinese origin were thenceforth classed as aliens, and were given stricter treatment before the law than the rest. This legal concept of "nationality" (Minseki) was unchanging for each individual over time.® Reflecting as it did an official system of ethnic discrimination, this scheme was maintained with some rigor by the police. The classification was extended into census practice, and the system of control later became the basis for collecting census and vital statistics data. It kept distinct four prin­ cipal groups of people—"Taiwanese" (or "Islanders," as they were called until 1939 in the statistics), Japanese, Koreans, and Mainland Chinese. 5 The Russo-Japanese War intervened in 1905, the World War in 1915. See Taeuber, Irene B., and Beal, Edwin G., Jr. (October, 1946), p. 9. β The privilege of adopting Japanese nationality was extended to natives of Taiwan only in later years, and was not widely used.

THE GROWTH OF POPULATION

By these actions the Japanese set apart from all others those who were descended from the original residents of Taiwan. This segre­ gation in the statistics extended from the most obvious differentia­ tion—that of Japanese nationals—down to fine distinctions within the Taiwanese population itself.T It was consistent in both the popu­ lation censuses and vital statistics, making it possible to amalgamate the two systems of record-keeping continuously as long as they were maintained. TABLE 1 TAIWANESE BORN OUTSIDE OF TAIWAN, 1920 AND 1930

Proportion of All Persons Who Have Been Born "Abroad," by Age and Sex C ensus of 1920 Male Female

All ages 0-9 10-19 20-29 30-39 40-49 50-59 60-69 70 and above

0.2 0.0 0.0 0.1 0.1 0.5 1.3 1.9 2.6

0.1 0.0 0.0 0.0 0.0 0.1 0.2 0.3 0.7

C ensus of 1930 Male Female

0.1 0.0 0.0 0.0 0.0 0.1 0.3 0.8 1.0

0.0 0.0 0.0 0.0 0.0 0.1 0.1 0.1 0.3

Sources: Census of 1920, Descriptive Report with Appendix, Table 79. Census of 1930, Statistical Tables, Table 55.

Regulating the conduct of "alien" Chinese nationals in Taiwan was not the sole purpose in establishing the legal status of residents. The authorities also set out to halt the further migration of people from the mainland. In this they were apparently successful. Tai­ wanese were required to verify their nationality by certain proofs, chief of which was evidence of having been born in Taiwan. Taking this as a criterion, there seems to have been little admixture from the mainland to the category of "Taiwanese." For example, tables showing Taiwanese who were born "abroad" were published with the censuses of 1920 and 1930. If the information was accurate, prac­ tically the entire population of Taiwanese born after 1895 was act This fine detail was confined to a slightly different classification, based on "race." It provided categories within the group with Taiwanese nationality—aborigines, Tai­ wanese of Fukien origin, and Taiwanese of Kwangtung origin. The two classifications were not precisely the same, the one based on "nationality" being the more consistent at different times.

THE GROWTH OF POPULATION

tually born in Taiwan (see Table i). Almost none under 25 years of age in 1920, or under 35 in 1930, were recorded as having been born outside. Some who were older at these two dates had been born in China, but they were very few and were probably living in Taiwan at the time of annexation. These figures imply that migra­ tion after 1895 was probably confined to people who were appro­ priately registered by nationality, and counted as aliens in census enumerations.8 Hence the census category of "Taiwanese" was defined so as to exclude migrants to and from the island. This is significant to know, for it means that Taiwanese counted in any census after 1905 were exactly the survivors of those who had been counted before (plus those who had been born in the meantime). It also means that the population of Taiwanese was for statistical purposes a "closed" popu­ lation—one which neither gained nor lost people by migration. All change in the size of the Taiwanese population was therefore due to "natural increase," or the balance between births and deaths. One can thus speak precisely of population "growth" in Taiwan, without reference to migration as a complicating factor. Half a Century of Rapid Growth

The total population, and the sub-total of Taiwanese, are shown in Table 2 as counted on October 1 of each census year, along with the annual geometric rates of growth that are implied at each intercensal stage. Every year showed an increase. The only setbacks were between 1915 and 1920, when the island suffered three near-epidemic attacks of disease. Even over this relatively poor span the average yearly rate of increase was close to one per cent. The original popu­ lation of 1895 was more than doubled by 1945; when the vital sta­ tistics were discontinued in 1943, there were nearly six million Tai­ wanese. 8 By evasion of the law some Chinese might conceivably have slipped into Taiwan and passed themselves off as Taiwanese. A further acquaintance with the regulatory system in later chapters, however, will make this appear most unlikely. There is evidence of another sort that the Taiwanese were not diminished by re­ movals to the mainland. Briefly, detailed testing of the validity of birth and death statistics has shown that practically all persons born in Taiwan were later accounted for in a census enumeration, or as registered deaths if they failed to survive. Practically all who were enumerated in one census were similarly accounted for at later dates. Dur­ ing the 1930's, Japanese permitted small numbers of Taiwanese to leave the island for special purposes. The effect of these departures was very small, and raises no difficulty in dealing with the data of earlier years. Thus, if there was emigration, it too was clandestine and very small.

THE GROWTH OF POPULATION TABLE 2 GROWTH OF POPULATION IN TAIWAN

Total Population and Taiwanese Enumerated at Successive Census Dates and Average Rates of Growth for Taiwanese

Y ear

Total Population (All Groups)

Taiwanese Population

1905

3,039,751

2,973,280

1915

3,479,922

3,325,755

1920

3,655,308

3,466,507

1925

3,993,408

3,775,288

Per Cent Annual (Geometric) Rate of Growth

1.12 0.83 1.71 2.22 1930f 1930|

4,218,663 4,592,537

4,313,681 2.48

1935

5,212,426

4,882,945

1940

5,872,084

5,510,259

2.42 2.42 1943

5,962,000*

" This figure represents the Taiwanese enumerated in 1940 plus their registered natural increase from the census of 1940 to the end of 1943. f Excluding Aborigine Territory. I Including Aborigine Territory. Sources: Statistical Summary, Table 57. Census of 19S0, Statistical Tables, Table 2.

These figures are significant also because of acceleration in the rate of growth, for it increased together with the number of people. Over the entire census period growth averaged 1.76 per cent per year. By 1940 Taiwanese were increasing at close to the rate of 2.5 per cent per annum. Compounded continuously, this latter rate would have enabled Taiwanese to increase twofold every 28 years, and twelve times in a full century. Natural increase of such dimensions is exceptional. Such growth in itself implies that some change has taken place in the organization of human activities. It necessarily reflects a situation that could not have continued long, and that cannot persist forever. In the world

THE GROWTH OF POPULATION

at large, for example, the increase of people that has populated most of the earth's surface has progressed steadily only in a few centuries of sweeping changes. It has maintained an average yearly rate of half of one per cent, only one-fifth that of the Taiwanese. At their more recent rate of increase, they could supply the entire world's present population in less than 250 years. Growth at this pace, for all its associated virtues in the saving of life and postponement of death, can be a costly way of expanding the membership of a society. It imposes on the adults the burden of supporting more dependent children than would otherwise be re­ quired. Consequently, at any particular time growing populations contain large proportions of young children and infants. The increase of Taiwanese left this kind of mark on their age com­ position. By 1940, they had passed through 50 years of growth almost unbroken by periods of high mortality. The age structure, both of males and of females, was one that expanded steadily at each lower age to a broad base of infants at the bottom (see Fig. 3). The pro­ gression was almost smooth, with few irregularities in between.

TAIWANESE POPULATION AGE

FEMALE

MALE

PER CENT

Γ Π 1905

B · 1940 OFFICE OF POPULATION RESEARCH, PRtNCETON UNIVERSITY

Fig. 3. Age Composition of the Taiwanese Population, 1905 and 1940.

14

THE GROWTH OF POPULATION

In 1905 there were only a little more than half as many Taiwanese as in 1940. Their distribution in different ages presumably had not been subject to this past influence of steady growth. It nevertheless had many of the indications of potential expansion. Births were at high levels in relation to the numbers of adults. Through the effects of high and erratic mortality, many never reached maturity. The inroads of past deaths had been far more irregular and gen­ erally more severe than in the population of 1940. As reflected in the dearth of persons from 5 to 20, for example, the turbulence that prevailed in Taiwan from 1890 to 1905 must have taken a heavy toll of infants and young children, and possibly prevented many others from being born. Even so, the population of Taiwanese in 1905 was an extremely youthful one. Provided that the supply of new infants each year would continue undiminished, it needed only an energetic set of measures for the control of mortality in order to produce the growth that ensued. Because in most census tables "nationality" was adopted to classify the population, the Taiwanese must be considered for most pur­ poses as a single group. It was not, however, a completely homo­ geneous category. It contained the remnants of the aboriginal in­ habitants, and people from many different parts of China. Those of Fukien origin were predominant; their ancestors had made up the largest portion of all the migrants to Taiwan. The majority who re­ ported ancestry in Kwangtung were descendants of the earlier Hakka migrants. These two principal classes of Chinese had begun to mix by intermarriage, which explains the lesser gain in those of Fukien origin by 1935. But during the great natural increase of Taiwanese they both grew almost evenly.9 The aborigines had ceased to be an important part of the popula­ tion by 1905. No longer able to contest the seizure of their lands and destruction of hunting grounds, they had retreated to the moun­ tains of Taiwan where most succeeded in avoiding Chinese influence. It took many years before they were finally subdued by the Japanese. The statistical information about the aborigines is incomplete and cannot be compared from one date to another, and we must all but ignore them in discussing the rest of the Taiwanese. Paradoxically, β Already by 1905 the boundaries between these two groups were less pronounced than they once had been. Separate tables of the population in that year according to prin­ cipal language spoken showed a certain amount of mixture between the two. The re­ markable fact is that they remained as distinct as they did.

THE GROWTH OF POPULATION

this small and obscure minority have been subjects for an immense amount of anthropological investigation during Japanese rule in Taiwan, while the peoples of Chinese origin (for whom accurate and precise information has been available these many years) have been almost entirely neglected. The increase of aborigines in census enumerations probably signifies nothing in the way of growth; it was not until the 1930's that they were brought fully within the scope of the census. TABLE 3 ETHNIC COMPOSITION OF THE POPULATION OF TAIWAN Censuses of 1905 and 1935 1935

1905 Per Cent Persons of Total

Japanese 57,335 Korean" Taiwanese Fukien Origin 2,492,784 Kwangtung Origin 397,195 Other Chinese Origin 506 Aborigines 82,795 Others" 9,136

Mahs per 100 Females

Persons

Males PerCent per 100 of Total Females

1.9

152

270,674 1,479

5.2 0.0

116 60

82.0 13.1 0.0 2.7 0.3

112 108 218 96 1,757

3,939,966 735,334 265 207,900 56,808

75.6 14.1 0.0 4.0 1.1

103 103 109 101 217

β Koreans included with "Others" in 1905. Source: Census of 1935, Statistical Tables, Introduction, p. 32.

The numerical position of the Japanese community in Taiwan gives no clue to its strong influence in the course of events after 1895. Its growth was not of the same sort as that of Taiwanese. Few if any Japanese regularly lived in the island in 1895.10 By 1905 a settlement of 57,335 was recorded for the census, increasing to 270,584 in 1935 and 312,386 in 1940, or five times its earlier size.11 Japanese in Tai­ wan did not add to their numbers through natural increase alone, especially since their community was a highly masculine one (see Table 3).12 As a matter of fact, their registered natural increase from 10 One foreign writer at the time went so far as to assert that prior to 1895 "there was not in the whole island a single Japanese resident." See Davidson, James W. (1903), P- 592. n The military garrison was supposed to be excluded from all of these enumerations. 12 Marriages with Taiwanese were prohibited until 1932. Offspring of such illicit mixed unions were rarely if ever counted later as Japanese, making the shortage of women a serious restriction on the per capita reproduction of Japanese.

THE GROWTH OF POPULATION

1906 to 1940 amounted to only 146,597, and if accurate could account for only 57 per cent of the total gain. The numbers of Japanese in Taiwan therefore increased very largely by migration. As we shall see, this was a special type of migration, based on opportunities that did not prevail for others in Taiwan or for Japanese at home, and brought special kinds of migrants. By recruiting new members rela­ tively faster than the natural increase of the Taiwanese, they also raised their proportion in the total population (see Table 3). The only other non-Taiwanese group, excluding an insignificant number of Koreans, were the Chinese from the mainland. They are shown in Table 3 as "Aliens," mixed with a few other foreign nation­ alities. Numerically they were even smaller than the Japanese. Being the ones most likely to blur the ethnic boundaries, Chinese nationals were classified and marked out with great care from the rest of the population, as we have indicated. In census enumerations they would have been indistinguishable from Taiwanese, but for the necessity the Japanese felt to control their activities and forestall the develop­ ment of close ties between Taiwanese and the mainland.13 The Taiwanese, however, were the group of major consequence for demographic study. Up to 1945 they comprised practically the whole of the population taken at large. The proportion of nonTaiwanese never reached 7 per cent. Distributed in the total, all these others were minor elements; only when special considerations are introduced do they become important. In the following pages, attention is given to the Taiwanese population almost exclusively, for their development under Japanese rule, which produced this natural increase, is the main problem to be investigated. is [Taiwan.] Government of Formosa. The Special Population Census of Formosa, 1905. [English translation] p. 161.

CHAPTER II

"Economic Development" in Taiwan RULING TAIWAN was a task that the Japanese approached with high

hopes. Because their goals were ambitious, the problems were more formidable than they might have been with a more modest and con­ ventional colonial program. Their officials in Taiwan were the first overseas administrators in the Empire, and felt a responsibility to make the colony appear worth while at home.1 Whatever their wider aims, their strongest aspirations for the island expressed themselves in energetic economic policies—policies pursued with such determi­ nation that they profoundly altered many aspects of Taiwan's eco­ nomic life. In the course of their administration, Japanese nationals came to look upon enhanced production with a two-fold purpose: to secure profits for certain individuals and groups, and to promote the self-sufficiency and military strength of the Empire. Here we shall examine the techniques of their economic program, survey its operation, and endeavor to make some appraisal of the accomplishments. A closer view of its significance for the everyday affairs of the population is reserved for the next chapters. The situation that the Japanese first encountered in Taiwan was one of long-established tradition, in which a small amount of inno­ vation could go a long way. Agriculture was the main source of live­ lihood for practically all Taiwanese. The particular manner in which farming is organized in Taiwan tends, as it does in most of the Far East, to shield it from the intrusions of the modern world. Under the Manchu regime the Chinese inhabitants of Taiwan had de­ veloped farming along the familiar lines of their forebears on the China coast. Farms existed on a local village scale and were operated for local subsistence. Most of any exportable surplus of production was removed from the villages in the form of taxes or rent. What was left became absorbed in the costs of civil disturbance, misgovernment, and expansion of settlement, rather than going toward broad­ ening commercial or other opportunities of the cultivators themselves. Whatever may bring about such a situation, it is one that tends to ι Most of the later Japanese publications about Taiwan's future also stressed the island's potentialities as a base for southward expansion of the Empire. The grandeur of these proposals suggests that some of them may have been invented in search of favor for their authors more than in serious expectancy of being carried out.

"ECONOMIC DEVELOPMENT" IN TAIWAN

discourage innovation in agricultural technique, and to keep at low levels the output per person working in agriculture. An economy in this condition requires important modifications when "economic development" is desired. Farming must be made more productive, and radical progress is necessary in other pursuits. Barring some cur­ tailment of consumption, such a change always implies expanding production faster than what is immediately consumed. One of the keys to the Japanese colonial program in Taiwan, and to its demographic consequences, was how they set out to induce this saving. They chose the course of concentrating the additions to in­ come from production in a few corporate business structures, instead of relying on the savings of individuals in Taiwan. Presumably this has been a goal in the management of most colonies, since indigenous sources have rarely been relied upon for large-scale investments. Its corollary—policies to suppress gains in individual native incomes— has been the chief source of hostility against colonial powers. But while most other nations have had only moderate luck in pursuing this objective, with the Japanese it was an outstanding success. They took care that the formation of capital proceeded rapidly and from internal savings alone. Even in Japan, which in some re­ spects is closely comparable, the whole pattern and purpose were different. While the Japanese were leading their own economy to­ ward self-sufficiency in manufactured commodities, economic change in Taiwan was being contrived with a still surer hand to move in an opposite direction, toward complementarity with Japan. The output of the island was gradually shaped so as to furnish a net contribution of agricultural and semi-processed food and related commodities every year to Japan. These were items which Japanese felt to be essential for their more ambitious plans at home, and which for various reasons they wished to avoid purchasing outside Japan in ordinary world markets. All the while, they mounted an impressive program of building up physical equipment in the island itself. Thus earnings were being reinvested on a large scale at the same time that they were also being drained off at a considerable pace by exports to Japan. This feat had to be accomplished in only one way—by raising the output of the island sharply while not greatly changing the levels of consumption for the population at large.2 In this respect the Japanese managed 2 The alternative—o£ gains by depressing popular consumption—should probably be ruled out. It cannot be conclusively demonstrated either that levels of living rose or

"ECONOMIC DEVELOPMENT" IN TAIWAN

the process of capital formation in Taiwan in basically the same way as at home, by prodding the economy at critical points with innova­ tions and developing special social policies to divert the gains from consumption to investment. The success of domestic saving in Tai­ wan is simply accentuated by the steady flow of commodity ship­ ments to Japan, whereby Taiwan as an enterprise did more than pay its own way. This success is best summed up in the yearly figures of trade with the outside world. From 1909 on, with the sole exception of 1913, there was an export balance every year in the value of Taiwan's ex­ ternal trade. Most of its entire trade was carried on with Japan— over 90 per cent after 1930 if Manchuria and Korea are included.3 Related to the total value of exports, the balance amounted to a steady share of about 30 to 40 per cent of all exports to Japan, 25 to 30 per cent of all exports to the entire world, that were not paid for in the form of imports of commodities (see Table 4). It is not proper to infer from these figures that Taiwan occupied a "favorable" posi­ tion, since apparently the balance was merely carried on the books of a few companies with home offices in Japan. This being the case, such figures probably even understate the amounts taken out.4 The Role of Government Policy

By implication, the foregoing account reveals that colonial rule yielded quick and gratifying results in Taiwan. These were not that they fell for Taiwanese. It does seem certain that on the average they must have risen somewhat, but that they did not move very far one way or the other. Undoubtedly some Taiwanese who took part in the new developments became richer, and many who did not became poorer. Agricultural output per farmer definitely increased, as the next chapter will show. There are figures (see Ch'en Cheng-hsiang, 1950, p. 151) that sug­ gest a slight falling off of rice consumption per person, but they must be computed in a way that weakens their reliability. However, the large additions to Taiwan's economy under the Japanese were the new productive facilities, fixed equipment and market organization, which it is the purpose of this chapter to document. There is no way of showing how these improvements were shared by the Taiwanese, though we know that the Japanese took steps to keep the major shares to themselves. 3 It is fairer to consolidate the "Yen Bloc" as a whole in computing these balances. To some extent imports from Manchuria and Korea were used to offset the account with Japan; this combined account shows a smaller balance than the account with Japan alone in most years. Separate data for Manchuria were not available in preparing Table 4. Aside from their monetary value on a balance sheet, the articles imported from Manchuria and Korea had a special importance, for they consisted mainly of fertilizers that were indispensable to Taiwanese agriculture. 4 Grajdanzev (194a), p. 157, claims that the export balance cannot be materially re­ duced by taking account of hidden items; many of these, such as cash remittances to Japan, would actually increase it.

24,659 59,936 130,192 195,767 256,326 251,476 445,653 527,699

Total Exports

* Figures are yearly averages over each period, t (—)indicates an import balance. Source: Statistical Summary, Table 821.

1899-1909 1910-1914 1915-1919 1920-1924 1925-1929 1930-1934 1935-1939 1940-1942

Years

TABLE 4

13,932 46,727 98,929 163,566 214,533 230,579 402,641 419,570

Exports to Japan and Korea

56.5 78.0 76.0 83.6 83.7 91.7 90.3 79.5

Per Cent of All Exports Going to Japan and Korea

—l,574f 4,212 36,943 62,039 65,862 75,720 115,005 97,419

Amount

—6.4f 7.0 28.4 31.7 25.7 30.1 25.8 18.5

Per Cent of Balance to Total Exports

Balance of All Exports over All Imports

VALUE OF EXPORTS FROM TAIWAN, BY PLACE OF DESTINATION, 1899 TO 1942* (Thousands of Yen)

"ECONOMIC DEVELOPMENT" IN TAIWAN

achieved through the efforts of private individuals in a free invest­ ment market. The Government-General of Taiwan was given a free hand by Tokyo in most internal affairs, and displayed great interest and vigor in regulating all areas of business activity that it could effectively control. A large part in the formation of capital in Taiwan can be attributed to official steps deliberately adopted with this end in view. On the one hand, these measures consisted of widespread ration­ alizations in agriculture and large investments providing technical economies to individual farmers; on the other, there were various devices to centralize the heavy investments in large blocs and thereby keep control over the entire venture. The former consisted mainly of improvements that could be initiated and carried out by Japanese without full participation by Taiwanese, and the latter of ways to guide the process and garner the returns. The one served to raise the average productivity of working Taiwanese without depending on them for inception; the other to avoid dissipating the earnings of the colonial enterprise as higher incomes among the Taiwanese, by confining them in the charge of a few Japanese.5 Although the Japanese treated their domestic agriculture in much the same way, they made use of their own farmers' earnings in further­ ing the growth of industry. The fiscal role of rural Taiwan cor­ responded to that of rural Japan, but in Japan the several lines de­ veloped together and were complementary. Fewer opportunities of this nature were exploited in Taiwan, for the complementarities were all external. This is shown in the other side of the picture of trade with Japan—that of imports and their composition. More than half of the value of imports from Japan during later years represented items that were clearly for current consumption.® There was surely nothing like the emphasis on heavy capital goods among the re­ mainder that is found in the imports of Japan itself.7 Problems of Pacification

This study is concerned as much with broader aspects of organiza­ tion in economic life as with the economic activities themselves. Thus 5 One of the organizational devices for diverting this income away from consumption was, as in Japan, to make it flow into corporate rather than individual hands, where its disposition was more easily regulated. 6 I.e., articles such as food, clothing, paper, beverages, textiles, manufactured con­ sumer commodities, and the like. One suspects that a large part of these went to main­ tain the comforts of the Japanese living in Taiwan. 'For example, see Reubens, Edwin P. (1950), pp. 109-112.

"ECONOMIC DEVELOPMENT" IN TAIWAN

it is pertinent to examine the stages by which the Japanese spread their influence over business affairs. Before the Japanese occupation, Taiwan was a busy farm area producing chiefly rice and other food­ stuffs for itself and neighboring countries. It still held virgin lands that yielded well to the cultivation of Chinese farmers, and was known among Mainland Chinese as one of the few places where fortunes could be made or lost quickly. It was known elsewhere for other features not unrelated to these—foreigners established some renown for the island's production of marketable crops like tea and camphor, as well as for its beautiful scenery and adventure. Some early Taiwanese had always enjoyed years of prosperity, subject to the vicissitudes of weather,8 warfare among powerful families or with the aborigines, and to the unpredictable greed of officials. Although agriculture was the mainstay of livelihood, there were the subsidiary pursuits commonly found from place to place in China. Additional opportunities arose in connection with commerce, which was the most flourishing sphere of enterprise. The export trade in the few products shipped abroad was the only spot affected by foreign influence. A small number of Western traders had stayed in Taiwan, arranging for the sale and delivery of the island's produce to the outside world, but the scope of their business operations was limited to buying and shipping. Even tea-growing, which an Englishman had begun, was left entirely to the local population and Chinese middlemen, up to the stage of packing and sending the tea abroad. The settled portion of the island, often down to the actual ap­ pearance of its villages, was a close replica of conditions on the main­ land of South China, 100 miles away. The Manchu government adopted policies calculated to restrict rather than encourage any growth or diversification of economic activities. It abhorred the pros­ pect of having the profits of any sponsored enterprise go to foreign merchants, and suffered an incapacity for such sponsorship inherent in its organization.9 Observers of the period saw little likelihood of change. Hence Taiwan when the Japanese acquired it was already a gran­ ary ready to function as a going concern, but with some impediments to being operated as such. In contrast to both China and Japan, it β These hazards were many and frequent. Lack of systematic attention to irrigation had placed farmers at the mercy of floods, local droughts and destructive hurricanes, and severe earthquakes were common. 9 This aspect of early political and business affairs in Taiwan is described at great length in Davidson, James W. (1903), Chs. 14, 15, and 17.

"ECONOMIC DEVELOPMENT" IN TAIWAN

was a newly settled land whose surface still showed little strain of intensive clearing and cultivation, and whose manifest possibilities for wider exploitation in agriculture and other lines had scarcely been tapped. Merely imposing public order and effective political administration could hardly have failed to ensure general prosperity, and the Occidental community at the time contemplated the experi­ ment with great interest. The new authorities did not leave much room for doubt about their intentions for the economy. Even while engaged in some dis­ sension and maneuver for power between their own civil and mili­ tary officers, they quickly set about constructing railroads as a state enterprise, reinstituting an official monopoly of camphor (amounting to a world monopoly since Taiwan with Japan was the sole important world producer), nurturing one large company for practical research and experiment in sugar production, reviving the former domestic monopoly of salt, and executing a complete and accurate land sur­ vey—Taiwan's first—in order to regulate some of the discordant re­ lations of tenancy among farmers and landlords. Most of the administrative changes were based on a notorious need for reform of pre-existing practices. But all of these reforms were two-edged. Besides regularizing commercial practices and "correct­ ing abuses," the official monopolies and export associations slowly made it impossible for foreign concerns to remain in operation. The revenues from key lines of business soon became centralized in either the Government-General or corporations explicitly controlled by it. Those relations began to appear between state and business that de­ termined the future course of the fields involved. The land survey furnished a basis for levying a stiff land tax which brought forth a gratifying and dependable flow of revenue into the public treasury and helped to finance later development schemes. If the overall design of economic expansion had been fixed in the minds of influential Japanese even before they took possession of Taiwan, and to some extent this seems likely, its timetable was not. They did not foresee, for example, that almost ten years of costly and arduous military campaign would be needed to subdue the ab­ origines and thousands of Taiwanese in the lowlands and foothills who had reverted to banditry. And it became apparent only after many Japanese lives had been forfeited to sickness that this "pacifica­ tion" of the island had to be accompanied by full-fledged warfare against disease and a general sanitation program, in order that enough

"ECONOMIC DEVELOPMENT" IN TAIWAN

Japanese government officers could be stationed over the island to establish political order firmly. This delay in extending their mastery over the whole of Taiwan was a disappointment to the Japanese. It spurred them to concen­ trate their efforts at organization on measures other than force to bring the inhabitants under control. Once set in motion after some three years' experiment and delay, their methods brought swift re­ sults. They encircled and isolated the aboriginal tribes in the hills, and introduced a refinement of old Chinese forms of keeping local order—the so-called pao-chia system.10 Private Corporations and Japanese Investment Other measures followed to make the Japanese rule more effec­ tive. As the exercise of authority became more a matter of routine, the policy turned more and more to exploring improvements and enlargements of production. This progress in government can be traced in the annals of specific achievements, like the drive to raise rice yields11 and the public health program. By 1905 the Govern­ ment-General felt itself well enough established to take a detailed census of the population; the descriptive report of its procedures is the best evidence of a firm degree of control and competence in per­ forming complicated administrative tasks in Taiwan. For the next 25 years or more, until they began to shift their energies to preparation for World War II, the Japanese proceeded to work out their policies for Taiwan (outside of the Aborigine Ter­ ritory) with scarcely a serious check.12 This long interlude between two periods of tight regulation is sometimes known as one of official "liberalism" or laissez-faire toward business affairs. It can be so designated, however, only in contrast to the strictness that preceded and followed. In steady succession the government completed several major projects. It finished the land survey of the entire island, essen­ tial to all their later engineering works as well as to the desired 10 The pao-chia system was a scheme for assigning joint responsibility to village and familial groups for the conduct of their members. Its operation and the reasons for its effectiveness are taken up in the following chapter. 11 Mendiola, Ν. B. (1949), p. 76 ff. 12 The Taiwanese efforts at insurrection, as for example recounted in OPNAV, Civil Affairs Handbook (1944), pp. 173-177, tended to be exaggerated by their instigators. Due to their own clumsiness, these were quickly put down by the Japanese, though not without some local disruption. Inside the Aborigine Territory, on the other hand, the Japanese devoted many years of hard fighting and frustration to "pacification," and did not hold the entire island securely until the mkl-1920's, nor peacefully until almost 1940.

"ECONOMIC DEVELOPMENT" IN TAIWAN

supervision of agriculture. Through its own revenues the govern­ ment became independent of the annual subsidy from Tokyo with which it had been launched. By direct action and through state-owned companies, it enlarged communications facilities into a thorough public network of roads, bridges, railway lines, telegraph and telephone installations, tunnels and mountain trails; constructed harbor equipment to accommodate an ever-growing volume of shipping: expended a tremendous effort (if not large sums) on sanitation and disease control; relieved much of the uncertainty of weather for crops by an irrigation system which also reduced the destruction of floods. Finally, and in some respects the central note of the entire program, the authorities gave continual attention to ways of rationalizing agricultural production. It must be admitted that most of their techniques of agrarian im­ provement were relatively well known even at the beginning. These were already part of the world accumulation of knowledge by the time the program was well under way. The Japanese did not discover new principles of agronomy in Taiwan, but drew upon this fund of information with alert but minor adaptations to fit the case at hand. The rice yields from the land, for example, were increased spec­ tacularly through the spread of a new strain of seed, introduced in 1922 and cultivated thereafter under supervision by Japanese.13 Practices like breeding suitable varieties of seed, using fertilizer systematically to replace elements removed from the soil, rotating crops, and matching irrigation schedules to the growth cycles of plants have for many years been commonplace in the world. But they do not by magical properties overcome the handicaps of unenlight­ ened farming practices. Japanese characteristically did not pin all their faith on the ability of these methods to work out improvement in yields unaided. They painstakingly supervised the application of techniques, and coordinated the use of new procedures with judi­ cious—never extravagant—mixtures of capital. The most crucial capital input of the Japanese in agriculture was the administrative framework through which these technical changes were introduced. It was a channel of influence that they turned to good account with a cheap outlay of resources. Japanese themselves were not cultivators. They operated experimental farms at many places to study means of improving agriculture, always from the standpoint of what was practical in a land of peasant farmers. They 13 Mendjola (1949), p. 76 ff.

"ECONOMIC DEVELOPMENT" IN TAIWAN

also maintained a zealous staff of workers to put new findings into effect. Taiwanese villagers were mobilized into compulsory associa­ tions to utilize the information, and the program was carried on with close cooperation by local authorities. In addition to these measures, certain of the large Japanese ex­ penditures on outright heavy investment helped to modify agricul­ tural productivity of Taiwanese. The building of railroads, largescale facilities for irrigation and flood control, and in later years factories for processing agricultural produce all worked some radical changes in the relations between the farms and the markets abroad. But these reached down only occasionally to disturb the operation of individual peasant farming. What one Japanese writer has called the process of "capitalist development in Taiwanese agriculture" was not so much a matter of developing investment in agriculture itself as of surrounding agriculture with capital improvements.14 Many improvements were introduced directly by the GovernmentGeneral. Usually a charge was levied on their use or the expense was defrayed by heavy taxation. The government matched each of these steps of its own by prompting the large private firms to carry on parallel works inside their corporate domains. This sort of extra­ curricular investment was widely undertaken by Japanese companies, and led to the growth of giant integrated concerns, some larger than their counterparts in Japan.15 A remarkable set of inducements existed for engaging in this sort of corporate expansion. Taiwan in the first place had the reputation of being the most profitable locale for investment in the Empire.16 In addition to this, the Government-General frequently offered sub­ sidies, partial subscription of funds by itself, special legal privileges, a trifling burden of taxation, and in general more cordial relations with the state than were available elsewhere. The response to these terms was apparently more enthusiastic than had been anticipated. It rose at times to feverish speculation, followed by short depressions among Japanese financial circles,17 and led the Government-General to impose restrictive schemes for fixing prices and output. "Yanaihara, Tadao (1929), p. 140 ff. He points out that most Japanese held land as landlords, like the Taiwanese, not as capitalists. The exception to this rule—and it is a large exception—was the management of sugar-growing plantations by a few of the large sugar-processing companies. 15 Grajdanzev, A. J. (1942), p. 96. 16Yanaihara (1929), p. 146. 17 Price movements of rice and sugar in Japan intensified some of these efEects, espe­ cially in the late 1920's and for several years after 1930. See Grajdanzev (1942), pp. 99-101.

"ECONOMIC DEVELOPMENT" IN TAIWAN

There was still another category, a somewhat subtler one, by which Japanese investment was injected into Taiwan's economy. This con­ sists of what might be called organizational changes, like banking facilities and the formation of stable domestic markets for disposing of produce. Many such features had been notably absent before 1895, when most commodity markets were no more than local in scope, prices fluctuated widely, interest rates were usurious, and speculation was the principal business activity. This channel of influence was often not actually separate from others; some efficiencies of this sort appeared even as by-products of other more deliberate innovations. These advantages tended to flow from relations among the modern types of business organization. Hence such benefits went to Japanese businessmen, though they often opened up subsidiary avenues to wealth for a number of Taiwanese in commerce and in service en­ terprises.18 With no significant exceptions, the growth and spread of Japanese business activity in Taiwan served primarily to introduce capital into agriculture without greatly changing the operation of individual farms. It furnished a pattern of services between the fields and the markets in Japan through which to reap the economies of modern organization, and yet ensured that most of the proceeds were reinvest­ ed on the island in a pattern determined by official decision. The Preparations for War

During the early or middle 1930's the Japanese Empire began to turn toward preparations for war. The colonies were called upon to increase the measure of their contribution in other lines of produc­ tion besides Empire self-sufficiency in food. Taiwan was not abun­ dantly endowed with the heavy stuff with which the war was to be made—metallic ores were lacking, and coal deposits were not plenti­ ful—and so it was permitted to assume some of the burden of process­ ing materials formerly supplied to Japan as agricultural and other primary produce. This decision did not issue entirely from military or strategic considerations, since in many cases it amounted to no 18 Chartered banks were an exception. They harnessed funds for industrial investment, and were all owned and operated by Japanese. ". . . modern banks in Taiwan help to collect money from the population and loan it at a relatively low rate of interest to the big Japanese firms, while the Chinese [Taiwanese] have to apply for loans to other sources where they pay two to five times as high a rate of interest as that which the Japanese firms pay." Grajdanzev (1942), p. 126.

"ECONOMIC DEVELOPMENT" IN TAIWAN

more than following advantages which had been obscured by the official emphasis on foodstuffs.19 A general survey of Taiwan's part in the Japanese war program is impossible to undertake here. It is nonetheless pertinent to indicate what were some of the implications of this shift for the internal ad­ ministration and economic affairs of the island. The turning point cannot be dated exactly, for the decisions behind it were probably not made at any one point of time. They rather reflected a gradual reorientation of official opinion regarding the course of development that was desirable. From the early years of its existence, the Government-General of Taiwan had sponsored many special corporations to finance and help to organize a limited modification of agriculture. It made more intensive use of them during the 1930's to enlarge new industries that it wanted to foster. Bank credit was made available for new in­ vestments at low fixed interest rates. Subsidies were granted widely to induce old firms to enter new fields. Corporations were spared the burden of taxes, government revenue being made up through an ingenious set of other devices.20 Under exclusive privileges, Japanese companies were attracted to exploit rapidly many of the virgin min­ eral and forest resources of the island owned by the government. Business activity was hedged about with countless regulations in­ tended to invite mergers and reduce rivalries between firms.21 The sugar industry was the outstanding example of government indulgence toward private corporations and its effects. A few Jap­ anese sugar companies had made an early start, before World War I. These were favored with direct subsidies, marketing advantages in Japan and imported equipment far superior to that traditionally used by Taiwanese.22 By keeping and improving this initial advan­ tage, Japanese producers all but drove Taiwanese competitors out of the field. Some of them grew so quickly that they were under some compulsion to branch out into providing transportation and related Special advances were made in starting hydro-electric power output and aluminum and chemical production, however, which would probably not have progressed in any such manner except by dictation from high places. 20 For sources of government revenue, see Grajdanzev -(1942), pp. 133-137, and Sta­ tistical Summary, Tables 335-345· 21 Stock dividends of corporations, for example, were limited by law. Since profits were high, large sums were forced into business expansion. 2 2 An even greater prop to the industry as a whole was tariff protection in the Jap­ anese market, whereby Japanese consumers supplied the greatest subsidy of all. Here also Japanese companies were enabled to reap the major share of the benefit, for reasons given above.

"ECONOMIC DEVELOPMENT" IN TAIWAN

services for themselves, in order to keep output in pace with their own milling capacities.23 The industry had expanded, in fact, too rapidly for its own good. It had suffered several crises, brought on by fluctuations in the demand for sugar in Japan, where its tariff prefer­ ence led it to sell.24 To "stabilize production," the important pro­ ducers organized a cartel (under the Japan Sugar Association) with government permission and guidance. Buying of cane from farmers had from the start been controlled by agreements among processing firms; the cartel opened a concerted effort to prevent cane prices from rising. The need for this step arose from the fact that sugar cane and rice both were grown in the same soils and more or less during the same seasons. In effect, the two crops competed for land. Japanese had always complained that the Taiwanese had a perverse tendency to cultivate rice when it was more profitable. Under the cartel agreements, sugar firms avoided competitive buying by strict agreements to allocate areas of cane purchase. Owing to their size and diversity, they were able to per­ suade many farmers to grow sugar instead of rice by withholding ir­ rigation water from rice fields, and by making conditional cash ad­ vances to individual families. By such means the sugar cartel accomplished what the government could never have done unaided: it ensured a steady output of sugar at fixed prices for which it would have been irrational for individual farmers to produce cane.25 Sugar acquired a place of high wartime priority as a source for industrial alcohol, and beginning in 1939 the Government-General assisted sugar concerns to construct distilleries and make alcohol on the island. Under these circumstances it is not surprising that the sugar processing industry tended to pass into fewer and fewer hands. After a merger in 1940 more than 95 per cent of the sugar in Taiwan was milled by five giant companies. Though more extreme, the history of the sugar industry was essentially that of all large-scale Japanese enterprise in Taiwan. The Government-General itself began to promote industrial ex­ pansion during the 1930's by further participation in economic life. 23 Privately owned and operated railroads attained a length of over half as great as the entire state system. Not all were built by sugar concerns; the largest belonged to lumber companies. Most were of narrower gauge than the main rail lines. See Statistical Summary, Tables 440-442. 2* Costs were higher than in Java throughout the period, putting Taiwanese sugar at a serious handicap in markets outside Japan. 25 The Government-General did its part to encourage sugar cultivation by putting rice paddies in a higher tax category than comparable sugar fields. And in 1939 it simply took over the entire cartel by issuing a set of detailed regulations to dictate its affairs.

"ECONOMIC DEVELOPMENT" IN TAIWAN

First, it controlled certain key sources of income and authority, as before, through the legal monopolies. Net receipts from these enter­ prises (including salt, alcohol, camphor, tobacco and opium) made up one of the largest classes of revenue. Public railroads, irrigation and generation of electricity were also government monopolies—un­ official but nonetheless actual—which were operated more for the sake of the power that could be wielded through their services than for profit. They were developed so as to assert what amounted to administrative control over the private sphere of production. These were really little more than refinements of earlier devices. One of the important departures from previous government policy was to partake directly in more business activity. No data were ever published to show the extent of government investment side by side with that of private firms. It was thought to be at a level nearly as great as the whole private sector, but was mixed too thoroughly to be separated in any reliable fashion from the whole.26 Outright government holdings in corporations went beyond scat­ tered blocs of shares in individual companies. Many of these official firms were diversified and enlarged so as to absorb parts of related fields. For example, the larger electric power companies, owned or dominated by the Government-General, helped to build the nucleus of Taiwan's metallurgy industry. Most of the wartime aluminum production was carried out by such integrated concerns. The govern­ ment itself formed the Taiwan Development Company, in 1936, as a super-corporation to promote interrelated ventures of this nature. It speedily promoted a host of subsidiaries strategic for the war pro­ gram.27 Besides all of these extraordinary activities, the regular budg­ etary expenditures of the Government-General leaned increasingly toward injecting new state services into the economy, and probably contributed thereby to the inflation later in the war.28 Certain tables have been quoted many times that purport to show the amounts of these investments (see OPNAV, Civil Affairs Handbook, Taiwan—Economic Supple­ ment, 1944, pp. 97-98). We shall not reproduce them here, for none of the published data on corporate investment are very trustworthy. The publications never stated what criteria were imposed in their collection, and so it is impossible to know exactly what they represent in any given instance. They conceal the weight of government as op­ posed to private corporate holdings, and do not differentiate between large holding com­ panies and their subsidiaries. 27 Incomplete data on some of the known holdings of this and other of the larger firms are given in the OPNAV Civil Affairs Handbook, Taiwan—Economic Supplement, Appendix, p. 86 ff. Through omissions, these figures are known to understate grossly the extent of corporate concentration. 28 These services stretched from research grants to larger funds allocated directly for new subsidies of business. Ibid., p. 59.

"ECONOMIC DEVELOPMENT" IN TAIWAN

Industrial Growth during the War

After setting the stage for limited industrialization by their treat­ ment of large business in Taiwan, Japanese leaders after 1940 began actively to move the players. The centralized corporate structure had been expected to be adequate preparation for most of the demands of war. The Government-General made frantic last-minute attempts to expand specific fields in which Japan itself was short, particularly aluminum and coal. But before long the subordination of Taiwan's economy to heavy industry in Japan, which had seemed so satisfac­ tory, turned out to be a source of weakness. Naval losses started to interfere with shipping between Japan and Taiwan. Many of the small but essential items supplied from the home islands (special equipment, replacement parts, fuels, and skills) became unobtain­ able. Certain large imports such as fertilizer were cut off. None of these shortages was fundamentally important for the island itself, since food was sufficient for its population and little else had previously been imported. Lack of articles of the first sort did play havoc with some of the new industrial output, and caused much of the equipment to deteriorate. Exports of aluminum and coal slowed down to a trickle by 1944, and practically all shipping to Japan—even rice and sugar—was discontinued early in 1945.29 By the time of Japan's surrender much of the highly organized economic life had come to a stop, but Taiwan suffered no devastating damage as a result of the war. Agriculture was still conducted as usual, ex­ cept that without fertilizer and other accustomed capital the yields had fallen off substantially. Accomplishments of Japanese Colonial Policies

Having described the workings of colonial administration in Tai­ wan, one should have some idea of its accomplishments before going on to discuss in later chapters its consequences in the lives of the Taiwanese. For this purpose, the abundance of records kept by the Japanese is often a hindrance rather than a help. In general the eco­ nomic statistics are not up to the standards of the population data. Numerical tables are plentiful, but explanations of what the numbers mean are not. They disagree from source to source, seldom with any reason given for the discrepancies. One must therefore look for sum­ marizing measures, and not for fine detail. 29 See Cohen, J. (1949), pp. 154-156.

"ECONOMIC DEVELOPMENT" IN TAIWAN

The most efficient view of total changes in the island's economy is by reference to its statistics of trade. As indicated above (Table 4), Japan succeeded in securing practically all of Taiwan's external trade to itself, if exports to Korea are counted. During the 1930's even traffic with China, just across the Straits, was almost eliminated. Compared with a wide range of other Far Eastern countries, Tai­ wan under the Japanese came to hold a surprisingly high rank in the value of trade per person. At its peak shortly before 1940, the total per capita value of its trade was well above even that of Japan (see Table 5). Since this rank was attained under peculiar conditions, TABLE 5 ESTIMATED PER CAPITA VALUE OF FOREIGN TRADE OF FAR EASTERN COUNTRIES CIRCA 1939° (Expressed in U.S. Dollar Equivalents) Country

New Zealand British Malaya Australia TAIWAN Korea Japan Proper Philippine Islands Burma Netherlands East Indies Thailand French Indo-China China

Year

1939 1939 1938-39 1937 1939 1939 1939 1938-39 1939 1938-39 1937 1939

Imports

109 61 67 16 15 10 8 5 4 4 3 0.85

Exports

127 72 64 23 11 13 10 12 6 5 4 0.25

Total

236 133 131 39 26 23 18 17 10 9 7 1

0 As reproduced in OPNAV, Civil Affairs Handbook. Taiwan—Economic Supple­ ment (1944), p. 73. Source: Far Eastern Unit, Bureau of Foreign and Domestic Commerce, U.S. De­ partment of Commerce.

however, it does not mark Taiwan as an especially "advanced" area (in the sense of supporting bountiful levels of living) except by ref­ erence to the Japanese residents in the island. Taiwan's position in trade with the outside world was created by heavy and increasing exports of a few products, mainly crops sent abroad in only slightly processed state. Compared with what took place in Taiwan's business and financial communities, there was re­ markably little shift in the composition of its agricultural output

"ECONOMIC DEVELOPMENT" IN TAIWAN

(see Fig. 4). Together sugar and rice never lost their dominant posi­ tion in the value of all crops produced. Despite some greater em­ phasis on finished sugar as a source of alcohol, the cane harvests gained very little in value over rice, even when the expansion of manufactures after 1935 would seem to have made this likely. In these terms sweet potatoes, which were always prominent in the Tai­ wanese diet, kept nearly a constant share of the total.

VALUE OF PRINCIPAL CROPS, TAIWAN

100 •OTHER

-ANIMAL PRODUCTS SWEET POTATOES SUGAR

RICE

1930

1945

or net Of FOPUl ATIOW "CSCAHCM. »*i»tcCT0* UBIVf BtlTt Fig. 4. Composition of the Gross Value of Total Agricultural Production, Principal Crops of Taiwan, 1915-1945.

Japanese promoted certain specialty crops, such as bananas and pineapples. These never occupied a large share in the total output of agriculture, but they more than kept pace with trends in the major crops (accounting for some of the change in the proportion of "other" products in Fig. 4). Together with tea, these crops made use of land not suitable for rice or sugar, and became important supplements to the total agricultural output. The Japanese encouraged such prod­ ucts by providing assured markets and technical supervision in their cultivation, and built canneries to process pineapples for export.

"ECONOMIC DEVELOPMENT" IN TAIWAN

The various staple and other crops that made up the bulk of Tai­ wan's agriculture eventually became distributed in a stable spatial pattern over the island. Sugar predominated over a large area in the Southwest, where it displaced rice in most of the land within the major public irrigation system. Tea in the North, and other specialty crops in the center, occupied some of the inland and foothill sec­ tions.30 But it is important to note that rice culture, particularly in small paddy fields, overlapped with all other crops that were raised on the broad plain of Taiwan, and the labor-intensive and tradi­ tional farming practices associated with it have continued to be spread throughout Taiwanese agriculture down to the present.31 The exports of staple crops would not have risen as they did un­ less the yields of crops did the same. Gains in rice yields from the land did not come at once. They followed almost 20 years of pains­ taking experiment, starting in 1906, to find a strain of seed that could thrive under conditions in Taiwan. Almost no improvement was made until the so-called horai variety was successfully adapted from Japan. Following its introduction the average yields per hectare for all varieties gradually rose as much as one-fourth, primarily by growth of the area devoted to horai rice and the special techniques used in its cultivation. Neither the horai nor the native strains were greatly improved in yield thereafter; in fact, it appears that horai declined slightly from its high performance at the start as it spread to land less favorable than the experimental farms where it was first grown (see Table 6). Systematic irrigation was one of the measures that raised the yields of land. Between 1930 and 1936 a few projects were carried out, by the government and by private companies, to introduce controlled amounts of water into lands that previously had been cultivated as dry fields. To illustrate approximately the trend of these events, Fig. 5 shows the changes in the area of land in paddy cultivation and that in dry fields. (Paddy land does not correspond exactly with the land that is irrigated on a large scale, for there were many paddy fields not included in these projects.) Since the total area cultivated was in­ creased over these years, it is possible that the irrigation projects brought some land into cultivation that was formerly idle. More, however, was merely converted from dry to wet cultivation (see Fig· 5)· so See Ch'en Cheng-hsiang, Atlas (1950), Fig. 161. si Ibid., Fig. 70.

"ECONOMIC

DEVELOPMENT"

IN

TAIWAN

TABLE 6 IMPROVEMENT IN THE YIELDS OF RICE IN TAIWAN, 1916 TO 1945

Average Yields of Different Strains in Each Year, First and Second Crops Combined (Hectoliters of Rice per Hectare)0 Years

AU Rice

1916-20 1921-25 1926-30 1931-35 1936-40 1941-45

17.8 19.3 20.9 23.0 25.5 21.6

Horai Rice —

28.5 22.7 27.2 28.2 22.7

Other Per Cent of Rice Land Strains Devoted to Horai

20.1 19.3 20.4 21.3 23.2 19.7

—

4.6 19.9 33.7 47.3 59.8

9 Crop statistics in Taiwan were not kept in metric units. Figures in the Statistical Summary of Taiwan were all converted before publication. Source: Statistical Summary, Table 203.

Fig. 5. Area of Cultivated Land, Taiwan, 1910-1944.

36

"ECONOMIC DEVELOPMENT" IN TAIWAN

Land that was improved in this manner not only yielded more at each harvest: over a third of the paddy land yielded more than one harvest each year. Under schemes for rotation and inter-cropping devised by the Japanese, it was not unusual to take three separate harvests from many of these fields in a year. The labor of preparing the land, of weeding, and of harvesting was not saved by these changes. They may have involved some net additions—irrigation became a more complicated process in some cases, and the aqueducts for transporting the water required special maintenance. Japanese were aware that the soil could not be farmed more intensively without losing some of its fertility, and encouraged the use of commercial fertilizers, much of which they required to be prepared on the island. The remainder comprised the largest single category of Taiwan's capital imports, greater in value than the iron and steel and machinery which are sometimes identified with the needs of agrarian countries. Importing fertilizers for sale in Taiwan proved to be a profitable line of business; it was reported to be firmly in the hands of a few large Japanese concerns, a fact which should occasion no surprise. It was shortage of fertilizers that brought on declines in average yields of crops after 1940 (see Table 6, above). The shift in official plans for Taiwan after 1930 leads one to look for some indication of results in the composition of Taiwan's prod­ ucts. The search is not very rewarding. In terms of gross value of product, agriculture did not lose its place as the foremost line of pro­ duction until 1940 and later. Then it exchanged ranks with "In­ dustry"—because farm produce fell off momentarily rather than because of any sharp rise in manufacturing (see Table 7).32 The steady but slow decline in the share of agricultural production in earlier years had indicated a tendency to submit more of the agri­ cultural produce to processing before it left the island, and not a shift from farming to manufacturing. It is when the category of "industrial" production is examined in detail that some of the more recent trends emerge. In Table 8, evi­ dently industry in Taiwan had always been centered around food products. Until the mid-'thirties these had amounted to threequarters of the value of all manufactures. Only from then on did the effects of new economic goals begin to appear. The largest ad­ vances were made in some of the fields that had been smallest, espe32 The sharp rise in the numbers of Table 7 was largely a matter of currency inflation.

"ECONOMIC DEVELOPMENT" IN TAIWAN TABLE 7 GROSS RECORDED VALUE OF ALL PRINCIPAL TYPES OF PRODUCTION, 1915-1942

(Yearly Average for Each Period) Years

Total

Agriculture

A. Millions of Yen: 1915-19 262.7 1920-24 411.5 1925-29 559.0 1930-34 525.5 1935-39 901.0 1940-42 1388.4 B. Percentages of Total: 1915-19 100.0 1920-24 100.0 1925-29 100.0 1930-34 100.0 1935-39 100.0 1940-42 100.0

Fishing and Forestry

Mining

Industry

144.5 207.0 293.6 255.8 432.9 576.4

7.2 12.0 16.8 15.5 39.6 62.8

9.1 23.0 31.8 26.5 41.5 91.8

101.9 169.5 216.8 227.7 387.0 657.4

55.0 50.3 52.5 48.7 48.0 41.5

2.8 2.9 3.0 3.0 4.4 4.5

3.4 5.6 5.7 5.0 4.6 6.6

38.8 41.2 38.8 43.3 43.0 47.4

Source: Statistical Summary, Tables 225-229. TABLE 8 GROSS VALUE OF PRODUCTION RECORDED FOR "INDUSTRY" IN TAIWAN, BY LINE OF PRODUCT, 1921-1942

(Yearly Average for Each Period) Y ears

Total

A. Millions of Yen: 1921-24 165 1925-29 217 1930-34 228 1935-39 387 1940-42 664 B. Percentages of Total: 1921-24 100.0 1925-29 100.0 1930-34 100.0 1935-39 100.0 1940-42 100.0

Metal Machinery Textiles Products ir Equip. Misc.*

3 3 3 6 11 1.7 1.5 1.1 1.5 1.7

3 4 6 17 46 1.9 2.0 2.5 4.5 7.0

4 5 5 12 30 2.1 2.2 2.3 3.1 4.5

β Includes Printing, Wood Products, Ceramics, and Other. Source: Statistical Summary, Table 269 and Table 270.

Chemical Products

Food Products

20 32 29 45 91

13 20 18 38 80

122 153 167 269 406

11.9 14.7 12.9 11.6 13.7

7.9 9.3 7.8 9.9 12.1

74.6 70.3 73.3 69.4 61.0

"ECONOMIC DEVELOPMENT" IN TAIWAN

daily in chemical and metal manufacturing; their expansion did not have an appreciable effect in the production of all the island. In other words, in spite of its expansion, industrial production had only begun to make headway in relation to agricultural output. Before 1940, manufactured commodities were still subordinate to Taiwan's farm products (and did not amount to much beyond proc­ essed sugar). Afterwards the Japanese did not have enough time to marshal the island's resources effectively in industry. We know this was not because they failed to make an effort—it was because they started too late. Certain types of heavy investment had received their close atten­ tion ever since earlier years. Railway construction, for example, was pushed forward with enthusiasm from the very first. The time sched­ ule of progress in expanding the public railroads was far ahead of their actual utilization (see Fig. 6). This was an investment which GOVERNMENT RAILROADS,TAIWAN

1,200

1,000 OPERATING LENGTH (KILOMETERS)

eoo

600

400 FREIGHT (MILLION OF KILOMETER-TONS)

200

900

1910

1920

1930

1940

OFflCE OF POPULATION RESEARCH, PRINCETON UNIVERSITY

Fig. 6. Expansion and Use of Railroad Facilities in Taiwan, 1900-1945.

39

"ECONOMIC DEVELOPMENT" IN TAIWAN

the Japanese undertook primarily in anticipation of long-term ben­ efit, without being able to justify its initial cost for many years. In terms of kilometer-tons of freight carried, the railroad made only partial use of its physical facilities until late in the colony's existence.33 This sort of schedule was also followed in establishing a system of improved roads and trails over the island. In both cases, Japanese had to bear the burden of ambitious engineering projects when they could least afford it, for basic transportation networks are more valuable as networks than they are piecemeal. Public irrigation was a project closely tied to agriculture, and fol­ lowed a different timetable (see Fig. η). After 1895 there was no large IRRIGATED LANO S AGRICULTURAL POPULATION HECTARES, THOUSANDS

600

IRRIGATED LAND

500

PERSONS, MILLIONS

400

300

200

AGRICULTURAL POPULATION

100

1905

1910

1915

1920

1925

1930

1935

1940

O F F I C E O F POPULATION R E S E A R C H , PRINCETON UNIVERSITY

Fig. 7. Expansion of Irrigation and Agricultural Population in Taiwan, 1905-1943. 33 Then it suddenly became overtaxed with cargo connected with the war, stemming from industrial growth and from Taiwan's role as a military base. See OPNAV, Civil Affairs Handbook, Taiwan—Economic Supplement (1944), p. 60.

" E C O N O M I C DEVELOPMENT" IN TAIWAN

supply of unoccupied land in Taiwan that could be drawn into cul­ tivation, and so the needs of an expanding farming population and growing export demand had to be met by other means. The Jap­ anese turned again to large investment to improve the effectiveness of the land that there was. The manner of constructing and upkeep of irrigation canals in Taiwan depended on mass labor. They were turned into an investment of labor, employed so as to produce an important form of capital equipment. The extent of irrigation more than kept pace with rural population growth while the Japanese were in control. Water for this purpose became indispensable in maintaining the output of farms. It was a hidden element in the im­ provement of rice and sugar yields, for the new varieties of seed that were developed were types that needed to be carefully irrigated. Finally, the development of electric power was so closely related to the growth of Taiwan's newer industries that it should not be passed over without mention. Generation of electricity was one of the specialized heavy investments that formed the nucleus of the "in­ dustrial" progress after 1930. This is evident in a circumstantial way from the trends of capacity in generating power and the spread of factory employment (see Fig. 8). Factories in Taiwan came to depend on the cheap electricity that was developed at state instigation from water power. Judged from the size of their employment, the most important of the mechanized factories in the island were "food" (i.e., sugar) and metallurgical plants, which were the heaviest users of electricity. As the next chapter will indicate, the Japanese broadened their objectives in Taiwan too late to put its human resources to best use in the new effort. Drawing the developed wealth of the island to­ gether into centralized control had turned out to be singularly ad­ vantageous. It had erected an entire structure of corporate organiza­ tion from specialized financial services down to the coordinating of actual physical production of commodities. The Government-Gen­ eral had retained a close supervision over this structure, by licensing, regulation, and participation as a shareholder. As a program it quickly marshalled the production of Taiwan's farms into a profitable colo­ nial enterprise. When it came to pressing this development further, however, the Japanese made no real progress, for they had frustrated these aims in advance. Under agrarian conditions, industrial skills and experi­ ence in management are slower and more difficult to accumulate

"ECONOMIC DEVELOPMENT" IN TAIWAN INDICES OF ECONOMIC EXPANSION PERSONS, THOUSANDS

tooo k

uu.

350

140

300 -

120

FACTORY W O R K E R S — J / /L / / / / / FT

250 —

100

-I

200

80

/ r>

J / —J

150 -

60

r /

/

100 —

40

/

^ - E L E C T R I C POWER I CAPACITY ( C O N S T R U C T E D ) 50 -

/ ι

O 1910

1915

1920

1925

20

_/

1930

ι 1935

ι 1940

0 1945

OFFICE OF POPULATION RESEARCH PRtNCETON UNIVERSITY

Fig. 8. Expansion of Electric Power Capacity and Factory Employment in Taiwan,

1910-1945.

than material forms of capital. Japanese in the past had never cul­ tivated an adequate fund of these among the Taiwanese. The same mechanisms which had served so well in channeling a large part of the total product each year toward desired uses had suppressed any opportunities for industrial employment that might attract Tai­ wanese who had administrative skills. What the Japanese themselves referred to as the "industrialization of Taiwan" was nothing but a corporate superstructure placed on an agrarian base. This was not capable of diffusing an industrial culture thoroughly among the gen­ eral population, because it had been devised to prevent this very thing.

CHAPTER III

Economic Activities and Social Change THE activities, interests, and aspirations of the Japanese community in Taiwan were things that existed apart from the indigenous popu­ lation. These elements were never effectively communicated to the Taiwanese—partly through a design to withhold them, and partly through reluctance among the people of Taiwan to accept them. Consequently the conduct of Japanese residents is not of much help in understanding the accustomed modes of life among the island's people. It is the latter problem that has the more interest for demog­ raphy, and makes it essential to consider the background of social order in Taiwan and the changes under colonial rule. The Early Agrarian Order in Taiwan

The importance of agricultural pursuits in imposing a predomi­ nantly agrarian character on a society is not a thing that can be pre­ cisely measured. Short of perfect exactitude, a great deal can be learned with lesser devices, by describing the cultural setting of agri­ culture in Taiwan, and pointing out how it influenced the participa­ tion by Taiwanese in the island's economy. The basis of this setting lay in the conditions that prevailed before Taiwan became a colony of Japan. By the mid-igth century, the lowland region of Taiwan was rapidly becoming fully settled. Piracy, which had made Taiwan a scourge to sea traffic along the China coast, gradually subsided into sedentary agriculture. Most of the Chinese population was already supporting itself by farming and the marketing of farm products. Since 1700 immigration had brought in a steady stream of Chinese from the mainland. In spite of sporadic prohibitions, hundreds of thousands of peasant farmers came to seek refuge from rural poverty in South China. Once arrived, they often reverted to this same condition, by reaching unfavorable terms with some of their countrymen already entrenched there as land speculators. Poverty tended to reproduce itself among those who reached Tai­ wan, though there was also great wealth open to the few who rose from cultivation to commerce. Each settled family held an advantage over newcomers. This superior position placed new migrants in

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

some competition for the more desirable lands near the coast. In order to maintain their accustomed manner of farming, many were forced to move inland, where they gradually dispossessed aborigines and the earlier Hakka settlers and pushed eventually to the foothills of the interior. These many-sided conflicts of interest gave rise to land tenures more complicated than those usually found in China. The result was a set of property customs that was to become con­ fused and encumbering in later years.1 With agrarian land tenures, however, it is all too easy to mistake form for substance. The system of land tenures in Taiwan gained several features that are distinctive to itself, arising to incorporate new parties into an old system—the aborigines who first occupied the area, and Chinese who acquired land titles to profit from the arrival of immigrants. But, since these landed property relationships are of little interest to us from an antiquarian point of view, their lasting effects are more important: they contributed to a form of village organization that survived through the entire Japanese occu­ pation down to the present. The flow of migrants is remarkable in that it populated Taiwan, with little variation over the years, closely after the model of the South China coastal area.2 The resemblance to the mainland went beyond forms of settlement and land utilization. Politically, under Manchu control the island was supposed to be assimilated into the administration of the rest of China. It was governed as a regular political division—first as a prefecture of Fukien and later as a prov­ ince itself—by officials appointed directly from Peking. Taiwan inevitably presented special problems of control. In the first place, it could be separated from easy communication with the continent; eventually it grew almost as accessible to the Western powers as to Peking. The device most relied upon to tie the area to China was a weak one. The authorities tried, futilely, to isolate it from all external contact. From time to time, they went as far as to restrict the entry of Chinese from across the Straits, though even this had little success. ISee Ch'en Cheng-hsiang (1950), pp. 41-43, and Kawada, Shiro (1928). 2The instances of alleged "radical transformation" of land tenures (Kawada, Shiro [December iga8], p. 88) were primarily fiscal measures to reassert the central govern­ ment's share in revenues from the land. Such reforms often are superficial from the standpoint of village organization. By way of illustration, one of the early official acts of the Japanese in Taiwan was to sweep away a whole section of these archaic land titles. While these reforms cleared the way to regulate rural areas from outside, they scarcely made an impression on property relations within the villages.

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

Furthermore, higher officials were themselves aliens, responsible ultimately to the Manchu court. Their tenure depended on their standing in Peking. Special competence or energy in office was not an advantage—it could even be dangerous, by making them targets of rumor and innuendo. Careers were sometimes ruined on the sus­ picion that certain officials had separatist leanings or were friendly to foreigners. Personnel were transferred regularly, as in China, to remove them from temptations of familiarity with their offices. Their only safe path to success was to maintain stability in local affairs, and somehow to prevent civil discontent from turning into civil dis­ turbance. This system was calculated to engross its key officers in their per­ sonal relations with Peking. It proved to be more effective in ensur­ ing loyalty than in promoting careful government or preventing corruption. There were few deterrents to peculation in office, which was respectable if prudently pursued, and surrounded by many op­ portunities—such as legal monopolies and illicit export trade—not present in China itself. These limited and negative objectives of government invited con­ tinual lapses of internal rule. Being unfamiliar with local conditions, officials simply did not govern in rural villages. Throughout the countryside they kept the peace by intrigue with powerful families rather than through actual administration—families that relied heav­ ily on property obligations among the peasantry to uphold their influence. The situation put a strain on the administration's own underpinnings, and led to severity in abuses and civil strife to which the mainland itself was unaccustomed. It is a testimony to the service­ ability of traditional Chinese institutions that they survived this test. Whatever its internal shortcomings, this scheme of political order was well suited to rebuff foreigners, at least foreigners whose ad­ vances were not backed by force. Acute frustration among the West­ ern commercial agents in Taiwan stands as evidence that it was accomplishing this part of its purpose. They complained bitterly that officials had connived at Taiwanese violence against Europeans and had even incited such incidents.3 The best information available about Taiwan's external contacts is contained in the published reminiscences and travel accounts of foreign trading representatives. The literary interests of these men throw light on the main preoccupation of their trading activities. s For example, see Pickering, W. A. (1898), Ch. 17.

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

Though their books have a wealth of illuminating anecdotes about the export trade, they provide almost no observations of any value about the ordinary affairs of the population. One might imagine from reading such sources that everyone in Taiwan was engaged in foreign commerce or some form of government service.4 These informants were, up to a point, honest men. They were not trying to mislead, and deserve credit for considerable acuteness and sincerity. The most likely reason for their failure to inform us about the Taiwanese is that they never had a chance to find out for them­ selves. Western merchants simply had no occasion to form more than special business associations with Taiwanese. Practically all of their transactions were conducted by contract merchants ("compradors") as middlemen. Foreigners not infrequently made efforts to inter­ fere in internal affairs, but these all arose over specific issues. They were attempts to make the social order more nearly serve their ends, not to alter it. As in China, trading relations carried on in this manner did not further the exchange of ideas along with commodities. They were contrived so as to minimize such interchange. This mutual aloofness worked both ways: it enabled outsiders to get along in ignorance of conditions prevailing in the island, and at the same time shielded the general population from contact with the outside world. Foreign commerce, the foremost potential channel for innovation until the conquest by Japan, was severely circumscribed by seclusiveness of the parties on both sides. The reasons for this isolation are not all to be found in the hos­ tility of officials or in the provincialism of Westerners. The foreign­ ers were eager to obtrude their presence further into foreign trade, even if only on their own terms. Government officers were too inept as administrators, their conduct too deeply undermined by directives from above, to succeed in stifling any positive tendencies toward change if such had arisen. The real bulwark of conservatism behind 4 These authors made better traders than ethnographers. They often did not go be­ yond expressing impatience that Taiwanese should not follow 19th century English trading conventions. The best and most comprehensive of these works (Davidson, James W. [1905]), out of a total of 32 chapters, devoted only a single superficial one to "The Inhabitants of Formosa," where the author did little more than describe the aborigines. The apparent exceptions prove the rule. If good material about those Taiwanese of Chinese origin is scarce, books about the aborigines of the interior (of varying quality) are not. Few Westerners penetrated to the dangerous mountain regions of Taiwan, but all of them developed immense curiosity about the customs of these "savages," which was not held toward the Chinese. See, for example, the titles in Takekoshi, Yosaburo (1907), pp. 321-342.

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

this nominal government policy existed within Taiwan's rural vil­ lages. The bulk of the population was made up of peasant farmers and small tradesmen, who supplied the commodities for most of the export trade. The fact that their products were exported did not bring them into closer touch with outside markets. Farming and other village activities did not reach beyond local boundaries; the arrange­ ments for disposing of farm produce, as we have seen, were in dif­ ferent hands. The rural population led a settled existence. Taiwan being a small area, travel between places was not uncommon, but it was rare for families to change residence from one part to another. With poor roads and chronic banditry, even minor journeys entailed some hard­ ship and risk and so were undertaken chiefly for exceptional purposes. The same difficulties hindered the movement of commodities. In­ ternal markets had not been very effectively developed to transfer articles of ordinary use from town to town; reports tell of wide varia­ tion in prices for standard items such as food, and of prohibitive haz­ ards to products in transit.5 Often products could be shipped to the coast and on to China more easily than to points elsewhere in the island. Local stability fostered local solidarities. The inconvenience of communication helped rural settlements to become close-knit com­ munities, by segregating certain cleavages and antagonisms that the Chinese had brought with them from the mainland. Many had come from different localities in Fukien and Kwangtung, and had main­ tained their ancestral clan affiliations. Members of these groups fre­ quently settled together and did not intermarry with others. Even many years after annexation by Japan, the Japanese found that sev­ eral communities had kept themselves free of mixture with other groups. For these various reasons, village organization in many re­ spects became self-contained and self-sufficient.6 Property institutions helped to buttress the stability in these small communities. In Taiwan, as in China, land ownership was the only durable form that substantial wealth could take. It was durable ε Takekoshi, Yosaburo (1907), pp. 252 and 269-270. β The turbulent history of Taiwan has been carelessly represented as one of indis­ criminate civil disorder, its early population as "unruly." It is noteworthy that most of the disturbances apparently followed the lines of cleavage above. Far from portending a breakdown of tradition within the Chinese settlements, these incidents were contests of strength between distinct parties—powerful clans, government forces, and aboriginal tribes. None was directed against the form of the social order itself, and none ever threatened that order.

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

against climate and short-term deterioration in use, against destruc­ tion through violence, and against decline in the income that it pro­ duced. Since the farming population was increasing within a fairly definite area, it had additional value in speculation. Land was also relatively safe against confiscation by the government; not only could its ownership be hidden from casual inspection, but the government itself depended on the obligations owed to landlords by tenants to uphold its authority. Under agrarian conditions of farming on a small scale it is impossible to cultivate large holdings of land under a single management. Hence, the only way to amass large wealth is to become a landlord—beyond some point, necessarily an absentee landlord. In early Taiwan, considerable concentration of landholdings was reported. Ownership tended to gravitate into the control of large and powerful families, whose influence was the only recourse of government officials in asserting their rule.7 On the economic side, this concentration of wealth was not given a chance to develop into industrial-type investment. The landed gentry, often landed on considerable scale, were the only group of persons who had access to sufficient income to become capitalists.8 Practically, however, it proved more secure and profitable to let out their lands to tenant farmers. When they branched out into different enterprises, it was as an adjunct to their status as gentry. In South Taiwan, for example, " . . . sugar was manufactured from the cane grown upon the owners' plantations, or from that brought from elsewhere. The cane was grown by the tenants, or by those who through being heavily indebted to the landowner stood in the same relation as retainers . . . under it [this system] the master of the sugar factory secures a threefold profit. i. As landlord, he claims from his tenants from 15 per cent, to 20 per cent, of the cane as rent. 7 One might suspect that, as in China, members of these families were the officials at lower levels of administration. But there is no direct evidence on this point, aside from frequent references to the fact that clan solidarity, imported to Taiwan by the migrants from South China, was one of the few bases for organizing any kind of joint action by Taiwanese on a large scale. s The government was unable to fill such a role itself for reasons already noted. Officials, who were not permanent fixtures in Taiwan, confined their acquisitive en­ ergies to foreign commerce, where returns were quicker and they did not need to make long-term commitments. For example, it is significant that the government almost never operated its own monopolies, but let them out to many private agents and taxed the products as they were exported.

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

2. As capitalist, he secures 14 per cent, to 24 per cent, interest on all the money he may have loaned to his tenant. 3. In order to cover the running expenses of the sugar factory he, as master, appropriates half the cane brought in by his tenants. "These vast profits leave very little for the poor tenants, who have therefore no means of improving the soil, but fall deeper into debt, until at length they become mere slaves of the sugar factor. This system has proved the greatest hindrance to the growth of the business."9 Establishment of Japanese Authority

Japan's objectives in enlarging its Empire required a more ener­ getic administration. With the background of past failures, it is un­ derstandable that the Japanese who first arrived in Taiwan met diffi­ culty in trying to impose this kind of rule. It quickly became evident that neither profit nor political dominion was assured to them with­ out some extraordinary measures. Japanese first installed themselves in economic life exactly where outsiders had done so in the past—at the periphery of traditional marketing and property relations. They remained most active, as we have seen, outside the realm of agriculture proper. Through their steps in gaining access to the old agrarian institutions, we can now begin to see how far these changes impinged on agriculture itself. The Japanese did not proceed by sheer force, nor simply in a form of "economic penetration." The control was made effective by wed­ ding the two channels—business and government—and by manipu­ lating the pre-existing situation from a few vantage points at the out­ side. To make a success of this endeavor, the Japanese found it expedient to leave much of this rural social structure almost intact, and turn its basic features to their own account. Histories of Taiwan over this period usually emphasize only the opposition to the new government that was overt and armed. But it was not by military conquest that the Japanese eventually secured the lowland area of the island. The military aspect was over quickly. Organized resistance collapsed by the end of 1895. Within a year or so after formally taking possession, Japanese became aware that di­ rect attempts to alter the old rural state of affairs were too costly and slow, and were ineffectual against guerrilla bands that had wide9 Takekoshi, Yosaburo (1907), p. 243.

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

spread local support. Their program to govern the countryside di­ rectly was a recognized failure as early as 1898.10 Starting with the issuance of pao-chia regulations in that year, the policy was revised with a view to getting less ambitious results, but more quickly. These were rules of conduct imposed on inhabitants of villages and smaller towns, under the threat of punitive action against the whole community. Virtually all Taiwanese outside the Aborigine Territory (with special provisions applying to certain cities) were assigned to household groups (chia), which were in turn grouped into village units (pao) for supervision. Responsibility for law observance and public order was allotted jointly among the house­ holds of each village. Positions of authority, carrying explicitly de­ fined duties, were conferred on heads of prominent (i.e., wealthy and influential) families. The system was based on a shrewd practical knowledge of those customary obligations that were most binding among Taiwanese. It was a move toward fitting the treatment of opposition to its roots instead of to its symptoms. By making the regulation of villages fol­ low lines of kinship and property, these measures in effect were in­ sinuated into the indigenous social structure of the population. The rules placed a multitude of specific duties upon each chia as a group. The chia, through its head, was responsible for reporting all events such as births and deaths occurring within it, for reporting and controlling all movements of persons in and out of its area along with the conduct of the permanent residents; it put into effect the earnest health and sanitation rules that were often promulgated; it mobilized labor from its ranks for road-building and irrigation proj­ ects as directed from above; it was required to disseminate instruc­ tions for crop improvement and soil maintenance, bearing its own expenses; it supplied men for auxiliary police service, supported its own school facilities, assessed fines among its members, collected many taxes, and could be called upon to enforce additional rules decreed by prefectural governors.11 These latter features in fact proved so attractive in avoiding the costs of more formal adminis­ tration that the Japanese kept the regulations in full force through­ out the entire span of their rule. 10

Yanaihara, Tadao (1929), pp. 220-223. See OPNAV, Civil Affairs Handbook, Takao Province (1944), pp. 49-51; also Hsu Tzu-wei (1946), p. 160 ff. This latter author calls the pao-chia system "the most effective instrument for destroying organized Chinese resistance to the Japanese."

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

The arrangement had more than savings in the expenses of civil government to recommend it. It gave the new regime a channel of authority reaching into every village, however small. Perceiving the value of this, the Japanese took care not to let the system lapse through neglect. They kept a close watch over its functioning by means of a large and reliable force of their own police. More impor­ tant, by insisting on rigorous enforcement of the regulations, they launched the pao-chia organization into full-scale and sustained op­ eration from the beginning.12 With the supervision of a trusted con­ stabulary, it functioned through the medium of prior obligations and loyalties that were already present in every community, with­ out the need to nurture affection for the central regime among the population. Such a system—when it works—is a most efficient way of main­ taining peace and order in a society already strongly bound by local custom. The Japanese originally adopted it as an improvement over the use of force, but never abandoned coercion in its operation. They merely utilized the structure of pao and chia to apply sanctions against wrongdoing more selectively and less frequently. Penalties fell on those persons in positions to exercise a deterrent influence on others. As a further by-product, the refined coercive aspect helped to lessen the rewards that had to be granted for Taiwanese collabora­ tion. This was important, for in rural areas such a form of control lends itself readily to keeping the avenues to individual betterment within an accustomed scope.13 The impressive agricultural improve12 A system like this one can function only where there are strong local solidarities. It turns communities into separate compartments for administration. This fact may be responsible for some of its success in maintaining order, for it inhibits organized op­ position to established government from growing to any large scale. But the considerable success of the pao-chia system in Japanese hands cannot be accounted for by these divisive effects. They were also present in Nationalist and Imperial China, and there only helped to divide the society so far that no central administration was possible. The essential difference was that in Taiwan the Japanese adopted it as an opportunity not for avoiding local government, but for effectuating central control over local affairs. Besides using it to enforce a minimum degree of legal conformity, they intermixed a host of other functions already alluded to. In addition, the Japanese incorporated a regular scheme of favors to be bestowed on active collaborators, in the nature of edu­ cation, business opportunity, and appointment to government posts, and thus by-passed the usual effect of family connections in filling lower positions of authority. 13 In later years many educated Taiwanese protested, and correctly so, that it worked as a levelling mechanism to impede all of them from rising very far above the status of peasants. In order to serve a central government at all, and not be merely a loose alli­ ance of towns and villages, such a system must offer inducements to a gentry class for cooperation: this complaint by frustrated persons with schooling should properly be

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

ments, health program, and construction for public works all de­ pended on generous amounts of labor. Without such cheap incen­ tives to mobilize this labor, these projects could not have been pur­ sued so effectively. As obvious as these advantages of "indirect rule" may be, it should be as clear that it is a poor vehicle for introducing social change. The more it is used to wield authority, the stronger and more re­ sistant to change it can become; the more it is undermined by alter­ ing local relationships that sustain it, the less remains of its useful­ ness. Hence, from adoption of this scheme initially as an expedient to "pacify" the island, Japanese authorities acquired a positive inter­ est in preserving the same set of agrarian institutions that had existed in Taiwan before their arrival. Thus they abandoned the idea of reorganizing the rural society— primarily because they tried and failed, but also because as it stood it gave them the best access to regulation of the agricultural popu­ lation and farming practices. It was in some respects a ready-made instrument for exerting the necessary control with the least disturb­ ance of customary affairs, having in the past enabled the various parties to the export trade to prosper at the expense of farmers. For better coordination of their policies, Japanese gradually did erect a more formal type of government administration, but the basis of the system in rural areas was not greatly affected by this modification. The Impact of Improvements in Agriculture

The Japanese who were charged with remodeling Taiwan's econ­ omy gave first priority to long-range investments around the fringes of agriculture. The rural economy was already geared to exchange food and related products for special capital services. Certain of these services were soon altered under Japanese influence. Positions that provided them, or that could usurp them, gradually began to be occupied by Japanese nationals—official, private, and always an indistinct class in between that wielded quasi-official sanctions. Those innovations that worked their way into agriculture were directed chiefly to raising crop yields per unit of tilled land. Another avenue of increasing the total agricultural product was by extending construed as a lament that there were few opportunities for advancement outside the rural villages and towns dominated by petty gentry. Discontent from this source bore few signs of being the awakening of a popular movement for independence among the Taiwanese, as it is sometimes represented. See Hsu Tze-wei (1946), p. 161 ff.

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

the land area that yielded crops. From 34.2 per cent of all "level" land (under goo meters elevation) in 1910, this area cultivated was raised to 43.7 per cent by 1940. The effective crop area was extended still further by irrigation, terracing and careful crop rotation. Irriga­ tion converted many dry fields into paddy fields, as we found in the previous chapter. Improvement in irrigation made more of the paddy fields yield two or more crops each year (see Table 9). Evidently TABLE 9 SPREAD OF DOUBLE-CROPPING IN IRRIGATED FIELDS, 1917 TO 1945

Area of Paddy Fields That Was Double-Cropped in Taiwan (Thousands of Hectares) Yearly Average in Each Period Per Cent Double-Cropped:

Y ears

Total Cultivated Area

1917-1919 1920-1924 1925-1929 1930-1934 1935-1939 1940-1944

730 753 795 816 850 844

Area of of Total Area in Double-Cropped Cultivated Area Paddy Fields Paddy Fields

336 366 386 422 516 520

217 252 274 300 320 314

29.7 33.4 34.4 36.8 37.6 37.2

of Total Paddy Area

64.5 68.7 71.0 71.2 61.9 60.4

Sources: Statistical Summary, Table 196. Ch'en Cheng-hsiang, Land Utilization in Taiwan, Appendix, Table 2.

much of the advance in crop production took the form of increasing the number of harvests from land already cultivated, which in Tai­ wan was already the "highest in the [Japanese] Empire."14 Land was not the only element of agriculture that was put to harder use. With a few exceptions,15 cultivation continued to be carried on with little difference in ways of applying human labor to the land. Agricultural implements were improved, but remained similar in design; irrigation facilities opened new land, but their construction was itself labor-intensive; transplanting of rice became generally practised to save land, but it is inherently a very laborious process. One of the major problems of farmers in the Far East has long been that their work is distributed unevenly throughout the ι* Nasu, Shiroshi (1941), p. 96. is A few large-scale plantations experimented with the use of tractors in raising sugar cane.

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

year. Double cropping fields not only uses land more intensively, but does the same for labor by spreading the intervals of greatest effort over a longer part of the year17 (see Table 10). 16

TABLE 10 LABOR REQUIREMENTS FOR RICE CULTIVATION IN TAIWAN

Average Man-Hours Worked for One Chia of Ricee Type of Work

First Crop Man ManwithAnimal

Seed Bed 6.3 Preparing Land 22.6 Transplanting 11.1 Fertilizing 7.1 Weeding 29.6 Irrigating 9.6 Pest & Disease Control 1.0 Making Windbreaks 0.1 Harvesting & Threshing 25.9 Total 113.2

0.8 18.1 0.1 0.9 0.4 0.0 0.0 0.0 0.2 20.5

Second Crop Man ManwithAnimal

1.9 15.1 10.1 4.0 23.9 6.8 1.6 1.8 27.5 92.7

1.4 11.9 0.1 0.1 0.2 0.0 0.0 0.0 0.5 13.1f

9 Women's time was counted as 60-70 per cent of that of men, children's as 30 per cent. Figures represent averages from 36 farms, having a mean size of 1.6 chia. (1 chia = .96992 hectare) Table has no date. f Published total does not check with sum of column. Source: Mendiola, Ν. B. (1949), p. 72.

The data drawn upon thus far give no indication as to whether harder work and better techniques meant higher yields to farmers as well as to land. Fig. 9 helps in a rough way to settle the question. The number of farmers increased relatively more than the land area devoted to farms. It follows that the average gains in crop yields per person fell short of the gains in the yield of land. Fig. 9 also shows crudely the trends in the productivity of farmers, taking into account the major crops of Taiwan. They were unmistakably upward until 16 If small-scale household farms are subject to two or three peak periods of work each year, shortages of labor tend to appear seasonally. When the seasons coincide for all or many farms, the shortages become general and cannot be met by hiring workers, since other available manpower is also in great demand. To maintain a household large enough to cope satisfactorily with this problem, the goal cherished by most peasant families is usually to bear the burden of supporting members who may not be fully occupied for the rest of the year. " The net effect of double-cropping may even be to increase the amount of labor that goes into a given amount of output. However, some of the steps required to pro­ duce two crops where only one had been grown before are labor saving, i.e., involve the use of capital. The question of precise trends is too complicated to deal with here in these terms.

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

CROP PRODUCTION, TAIWAN

7 RICE,(HECTOLITERS) UJ O= D

5 3


943- 7° Per cent °f Taiwanese children of legal school age were attending some sort of school. In time this more liberal turn of policy might have paved the way for admitting Taiwanese into more responsible positions in the island's new industrial program.

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

TOTAL OCCUPIED MALES, BY OCCUPATION TAIWANESE

IOO

JAPANESE I N TAIWAN

I N JAPAN .OTHER GOVERNMENT, PROFESSIONS -TRANSP., COMMUNICATION

80-

COMMERCE

MINING, MANUFACTURING

40-

OFFlCE OF P O P U L A T I O N R E S E A R C H , P R I N C E T O N U N I V E R S I T Y

Fig. 11. Trends in Occupations in Taiwan and Japan, 1905-1940.

certain cogent advantages. The alien status of Japanese nationals was useful in keeping high standards of responsibility and perform­ ance among members of the public and private bureaucracies. This personnel could be interposed to administer Taiwanese affairs with­ out becoming susceptible to local pressures. Though the Japanese had gained access to the rural society by manipulating its traditional institutions, they had to guard against subversion of their admin­ istration by influence from the Taiwanese side. They did not succeed, as no large organization has, in making their officials completely disinterested and honest. There were scandals and reports of self-enrichment on the part of several of their highlyplaced nationals. When it came to collusion with Taiwanese, how­ ever, they were all but incorruptible. The instances of improbity that achieved publicity in Taiwan were mostly between government and business interests. Irregularities of this sort did not threaten the objectives of the regime, for these parties had more at stake together than at issue in the economic expansion. Much of what was lost in

ECONOMIC ACTIVITIES AND SOCIAL CHANGE

this wasteful scheme of recruiting from abroad therefore must have been regained in the form of security against the conservative in­ fluence of tradition from within. In one respect, however, this policy proved to be an unqualified loss to the Japanese. They established a caste-like status for their own nationals, hedged about with favoritism in employment, education, and business opportunity. This widened the gulf that the general severity of their rule had made between themselves and the Tai­ wanese. The enmity that they had earned among the population left them unprepared to bring it effectively into the last-minute indus­ trial effort called forth by the war.32 We know from the Japanese themselves that the latter 1930's were for them years of exceptional new decisions and vigorous efforts at expansion. Official policy began to assert itself more strongly in mat­ ters of industrial planning and investment. The previous success in aiding farm production, which had given official circles some occa­ sion for complacency, suddenly became inadequate. The govern­ ment began at a strenuous pace to mobilize Taiwan's economy for serving the broader purposes of the Empire. On very short notice they sought to transform Taiwan into a quasi-industrial appendage of Japan. The attempts to encourage new aptitudes and incentives among Taiwanese are as interesting to con­ template, in terms of their implications for manpower, as were the attempts to prevent this very development. The two objectives are of course incompatible, but they were fostered side by side by the Japanese under the stress of preparing for war. Yet by 1940, at the peak of all this energy, the portion of the working population em­ ployed outside of agriculture was scarcely greater than at earlier census dates—still only one-third, and this figure includes an unde­ termined number of Japanese residents (see Table 11). Only a tiny fraction of the people in Taiwan had taken up factory employment, and many of these were also Japanese. We are left with the conclu­ sion that Japanese colonial policies and indigenous tradition had combined to restrict social change within a small segment of the pop­ ulation. Japanese comprised most of this segment. New activities were spread thinly among Taiwanese, and left the remainder en­ gaged in the same pursuits as their ancestors for countless generations. 32 During the quickened prewar mobilization of their resources in Taiwan, despite the crowded condition of agriculture, Japanese businessmen actually complained of a shortage of labor in industry.

ECONOMIC ACTIVITIES AND SOCIAL CHANGE TABLE 16

Taiwan: Total Males, and Males with Occupation Distributed by Occupation, at Census Dates, 1905 to 1930. Total All Groups, Taiwanese, Japanese, and Others'* Occupation

TOTAL MALES Occupied

1930

1920

1915

1905

1,369,877

1,180,842

1,155,777

864,884 27,709 18,345 122,066 148,847 61,323 69,644 57,059

761,047 25,858 17,240 112,063 97,726 45,645 48,018 73,245

776,181 26,911 9,741 94,338 93,171 38,901 34,737 81,797

1,081,667 731,473 29,465 6,843 67,526 83,815 29,138 30,568 102,839

TAIWANESE MALES Occupied 1,265,185 Agr. and Forestry 860,613 Fishing (Excl. Salt) 25,916 Mining 16,970 Manufacturing 95,832 Commerce 129,372 Transportation 48,677 Gov't, and Prof. 34,260 Other 53,545

1,100,297 757,917 24,261 15,629 87,314 83,600 35,428 23,811 72,337

1,088,199 772,437 26,150 8,319 72,444 79,823 29,505 19,177 80,344

1,045,667 731,102 29,260 5,203 58,833 77,060 23,186 18,923 102,100

53,403 3,430 611 1,113 16,482 9,498 6,273 15,328 668

27,835 285 170 1,380 5,739 4,592 3,871 11,502 296

14,169 312 150 309

8,165 86 35 260

Agr. and Forestry Fishing (Excl. Salt) Mining Manufacturing Commerce Transportation Gov't, and Prof. Other

VPANESE MALES Occupied

75,812

Agr. and Forestry Fishing (Excl. Salt) Mining Manufacturing Commerce Transportation Gov't, and Prof. Other

3,474 1,617 411 13,796 11,043 8,360 34,511 2,600

63,853 2,815 1,310 1,187 17,530 9,567 7,466 23,751 227

>THER MALES Occupied Agr. and Forestry Fishing (Excl. Salt) Mining

28,880 797 176 964

16,629 315 280 424

ECONOMIC ACTIVITIES AND SOCIAL CHANGE TABLE 16 (continued) Occupation

Manufacturing Commerce Transportation Gov't, and Prof. Other

1930

12,438 8,432 4,286 873 914

1920

7,209 4,538 2,727 455 681

1915

5,410 3,849 3,122 232 785

1905

2,954 2,163 2,081 143 443

* For comparability, Domestic servants are excluded at all dates; Ceramics workers are as­ signed to Manufacturing in 1905 and 1915; Salt workers are assigned to Manufacturing from Fishing in 1905, 1915, and 1920. Sources: Census of 1905, General Volume, Table 25. Census of 1915, Statistical Tables, Table 66. Census of 1920, Descriptive Report, Appendix, Tables 51 and 52. Census of 1930, Statistical Tables, Total Island, Tables 45 and 46.

CHAPTER IV

The Disposition of Manpower THE human side of livelihood in the countries of Asia understand­ ably attracts serious attention. For many years both Westerners and Asians have displayed increasing interest in the tedium of work in this portion of the world. Human labor under these conditions has variously been held the object of pity for its drudgery, praise for its industriousness, impatience for its low productivity, and admiration for the contentment it sometimes fosters and for its integration with cultural traditions. Such impressions are not always a help in gaining a disinterested and factual grasp of the problem. They often are based on senti­ ment or feeling more than on actuality. Few of the commentators on the lot of the Asian peasant have ever come into contact with it at first hand, which may help to account for the diversity of views on the subject. Yet underlying all of them is a certain foundation of fact. No one will deny that work in these parts of the world is normally toilsome and arduous. It is, by our standards, conducted with such a lack of advantage from effective instruments that prodi­ gious amounts of labor must be expended simply to keep alive. In addition, as compared with industrial societies, a person's usual work is more firmly imbedded in the social position that he occupies, which in turn may bear less relation to the income that he gains or the commodity that he produces. No single project could properly cover all of these aspects of live­ lihood. For Taiwan the advantages of the data that exist lie in adopt­ ing a demographic view of work and discovering how laboring activi­ ties are distributed through the population. This approach does not enable us to answer all of the questions that usually arise about eco­ nomic activities and their setting, but we can move from it to a more adequate conception of the connections between livelihood and the social structure. This point of view was introduced in the previous chapter in order to fill out the one-sided picture that resulted from taking Japanese policy at face value and surveying the economy through its specialized corporate structures. Since the body of facts to be covered was large, the evidence of the occupational statistics has up to this point been treated cursorily.

THE DISPOSITION OF MANPOWER

Economic activity is usually approached in demographic analysis by drawing attention first to the portion o£ a population that is in­ volved. This group may be designated as manpower, or, more ambi­ tiously, as the "labor force." It consists of those persons whom the influence of customary rules, the demand for labor in the economy, and expectations of others have induced to work for a living. In this chapter we shall take up some problems of determining the nature and size of this group in a population such as that of Taiwan, and to discover in greater detail some of the patterns that existed in its composition. Since this inquiry is also a means to observe the insti­ tutional environment of work, we shall be concerned only with the people whose society is reflected in these patterns, the Taiwanese. The Nature of the Labor Force

By itself, the statement that the labor force is supposed to include people who work as distinct from others is not very useful. It is like knowing that humankind is composed of good- and bad-tempered persons—neither proposition gives any basis for assigning people to one class or the other. To classify a universe of individuals by any such principle requires a set of criteria for judging what activities should count as "work," and what degree of performance is neces­ sary to qualify a person as being fully employed. The first problem in specifying these criteria is a conceptual one: what purpose is the distinction intended to serve? This purpose may be to ascertain the usual state of affairs in a society where conditions of livelihood are relatively stable. One can expect that most of the population holds some employment status that is recognized in wellestablished usage. In this class the census procedure most likely to succeed is one that elicits a statement of the activities most commonly engaged in, according to this usage. Placing many arbitrary barriers in the way of such a response, like special rules for weighing levels of wages or exact units of working time, probably does not make the answers more precise. Such considerations may even obstruct the pur­ pose by calling for facts that are irrelevant to the situation.1 On the other hand, the classification may be designed as a basis for estimating the volume of employment at a particular time (or of "unemployment," in the sense of people not working who would if they could), mobility between different types of labor, or short-term ι cf. United Nations, Population Division and Statistical Office. Studies of Census Methods, No. 4 (1948), pp. 14-38.

THE DISPOSITION OF MANPOWER

fluctuations in employment. If employment is to be regarded in this way, it needs a set of criteria that are more complex, more sensitive to particular circumstances, and better oriented to the activities of people at a given moment. Census Procedures in Taiwan for Determining the Working Population

Even if the conceptual questions can be settled, there remains a group of procedural ones. Since information which might be used to measure the size of the labor force must cover the entire popula­ tion, it is ordinarily secured as part of a general census enumeration. Consequently most of the issues in collecting such materials are en­ countered in connection with census administration, and they have usually been dealt with by official organs charged with managing a census. Until very recently, it has been in this procedural realm that most attention has been given to labor force measurement—how to establish definitions, etc. In actual experience, some of these criteria of working status have always been incorporated into census data that are already collected and published. They are beyond the control of the persons who make use of the materials, who are as a rule not the same ones who gathered and compiled them. Most of them must be accepted as they are, save for slight alterations (like shifting or combining certain categories) which may or may not aid in comparing data from dif­ ferent sources. Often the best that can be done is to scrutinize care­ fully the standards that were used in any given instance, and to make whatever allowance is possible for peculiarities that may exist from case to case. In Taiwan, where census practices in Japan were followed,2 the first of the two concepts of employment was used with little variation from census to census. Some specific entry as to "principal occupa­ tion" or customary activity was required of each individual. This left it up to each person or household head to decide whether he held an occupation; if so, he was supposed to specify the usual activities that he actually performed, as distinct from the part of the economy in which the work was carried on.3 The information was compiled 2 The enumerations of 1905 and 1915 in Taiwan had been planned likewise to syn­ chronize with Japan in every major respect, but were carried out actually without the intended counterparts in Japan Proper. 3 That is, in the current census usage of the terms, the enumerations were "occupa­ tional" rather than "industrial." The compilation of occupational titles before 1930,

THE DISPOSITION OF MANPOWER

in detail, and transmitted to the central office, where it was reduced to groups of occupations by clerks. Here strict instructions were issued to ensure that the procedure of classification was uniform and consistent.4 The authorities were anxious to impress the outside world and their own superiors in Tokyo with their efficiency as a colonial gov­ ernment, and spared little trouble to prepare census reports that were documents of their competence. Thanks to elaborate and often ex­ aggerated care in these matters, there is some possibility of checking on the modes of allocating people among groups of occupations. Every occupational census was published with at least one detailed tabulation showing the separate occupations of which the broader categories were composed. In 1905, for example, the census recog­ nized 178 specific occupations. The number rose in the revised lists that appeared in later censuses, until some 376 were taken into ac­ count by 1930.5 In fact, the zeal to gain stature for their work along­ side the most modern Western enumerations carried Japanese census authorities still further. They prepared comprehensive glossaries of occupational titles to standardize the content of all categories that were tabulated. These translations contained up to several thousand names of occupations current in different dialects, and reconciled local usages with each other and with their conventional equivalents in Japanese. It is not so clear how persons who had occupations were distinwhile designed to classify occupational roles as such, tended to group together many persons with different functions who worked in certain lines. For example, workers in retail stores selling a particular article, regardless of whether they were owners, hired sales persons, peddlers, or simply helpers, were often placed in a single group. The cen­ sus of 1930 included a stricter classification of occupations, and introduced for the first time in Taiwan a separate "industrial" tabulation of occupied persons. With one or two exceptions explicitly noted, the former of these will be used throughout this study. It is the most nearly comparable from one time to another, and provides a more useful type of information for our purposes. 4For a short account of the administration of these measures in Japan in 1920 and 1930, see Taeuber, Irene B., and Beal, Edwin G., Jr. (1946), pp. 9-15. The execution of the census of 1905 is described in more detail in a descriptive report translated into English as The Special Population Census of Formosa, 1905 ([Taiwan], Government of Formosa, Tokyo, 1909). 5 The number of occupations included in any of these lists is in itself an important detail. Very likely the systems of classification were developed beyond the degree of specialization in the economy itself. This is especially true in the later years, for the occupational scheme was devised for use in Japan and transported bodily to Taiwan as part of the policy of coordinating their census practices. While useful in locating certain specific types of activities, these minutiae are relevant chiefly as evidence of the meticu­ lous attention to detail in the compilation of the materials.

THE DISPOSITION OF MANPOWER

guished from those who did not. There were few rules covering this point. Once a person reported some employment on his census sched­ ule, he could be assigned to an occupational class very simply by the central coding staff. But it was not easy, in terms of a "customary occupation," to control the basis on which a person might decide whether he was employed at all. The criteria for making this distinc­ tion ultimately rested on prevailing social standards of what consti­ tuted employment. No society is lacking in definition of its own of what is useful and what is not, though the particular objects of such evaluations may vary widely from place to place. But as surely as standards of useful­ ness are universal among different peoples, census agencies likewise will inevitably tamper with them in the course of their duties. There is some practical necessity for this. These cultural definitions are not often explicitly formulated, nor are they consistent for all individuals in a society. Both clarity and uniformity need to be stressed in an occupational census, and so certain conventional rules are always established. The virtue of a notion such as "customary" occupation in use—and its greatest weakness—is that it serves well with little overt supervision. Hence, while an occupational inventory along these lines can be a simple matter, some of its basic categories are necessarily indistinct. This type of inexactitude creates "fringes" of employment, espe­ cially among groups of people who are just entering or retiring from the labor force. Here are found the individuals whose exact occupa­ tional status would have to be decided on more explicit standards if all entries were to be made consistent. Customary occupations are generally suspected of being inflated, of retaining people who are temporarily out of work or who have retired from active work.6 The only procedure that census officials in Taiwan adopted to con­ trol this tendency was to specify carefully the circumstances for re­ porting no occupation, in order to settle the doubtful cases from this side. The "no-occupation" category thus had the most positive definition of all. All persons not reporting a regular occupation were required to state in detail their means of support.7 β Against this can possibly be set an honorific value, unusually strong in a Chinese society, that accrues to leisure activities, which may tend in the opposite direction. 11£ living on income, the type of income was to be stated; if a family worker, the type and family status was to be included; for "dependents"—those for whom no sep­ arate source of income could be found—it was to be noted whether students, family

THE DISPOSITION OF MANPOWER

In publication, except for some special tables on households, this information was usually lumped into one single class ("no-occupa­ tion"). The census of 1930 was the one exception, when the census authorities tried to be more precise about all occupations than before. In that year the following separate types of status were tabulated as sub-classes of the major group "without occupation": those living on securities, living on tenant fees, living on house and land rents; students, "dependents," inmates of special institutions, persons on relief, prisoners, others and those whose occupations were not re­ ported, and domestic servants.8 This policy towards the listing of all non-employed statuses should be kept in mind when interpreting occupational figures for Taiwan. It furnishes the only definite clue that we have of the boundary that was drawn to set off the labor force as a whole. The precautions taken to prevent the "no-occupation" category from being a repository for borderline cases may even have helped in exaggerating the amount of actual employment. Hence, while census data may give good evi­ dence on the kinds and patterns of employment in Taiwan, they are not free from bias in indicating the total amount.9 Variation by Age and Sex There is no way to clarify the meaning of data on occupations without investigating the purposes and problems of census-taking. But there are limitations. The outcome of these compilations de­ pends also on how well the planning and execution of the censuses were adapted to the circumstances they were supposed to describe. members, or other relationship to the household, and the occupation of the head of the supporting household. 8 It has been necessary to reassign domestic servants arbitrarily to the "no-occupation" class. So few were enumerated that the category is meaningless, especially with reference to males. β On conceptual grounds, there is not much reason to endow the "amount" of em­ ployment with much importance, for in demographic data one has to deal with units of persons and not of work. It is well known that a certain amount of temporary or partial unemployment is chronic in agrarian societies; it is taken for granted, expected, and institutionalized in some "hidden" form. Properly speaking, this is not unemployment at all: the persons involved would not necessarily work if work were available. In these circumstances the indistinct line separating these cases from the working force is not primarily a defect of the concept of "customary occupation," but reflects the facts of the real situation. Other uses of such data are not necessarily impaired by this circumstance. The bias toward over-stating the number of working persons did not vary greatly in Taiwan over time, and therefore will not affect comparisons of the same population at different dates, such as those used in this chapter and in Chapter III.

THE DISPOSITION OF MANPOWER

In general, activities having to do primarily with livelihood are so important in human societies that they are always subjected to some special forms of coordination. Never are they treated casually in a society, as minor points of etiquette or personal friendship may be. Much of their importance arises from the fact that the same per­ sons who engage in them also occupy other positions of social sig­ nificance. This coordination, therefore, cannot be primarily a matter of regulation within the realm of economic behavior,10 but consists of keeping such behavior consistent with the social order. Demographically, the problem is found wherever occupational functions overlap with other roles in the same individuals. Two of the commonest regions of overlap in a population are with sex and age. Most of Asia, for example, is beset by such hazards to human existence that its peoples have maintained their numbers only by bringing forth abundant offspring. High fertility, which until recent centuries had been the universal offset to high mortality, has always necessitated a distinct kind of social position for women. Be­ ing the only ones capable of parturition, females in these societies come under the influence of family institutions that attach high priority and emphasis to the bearing and rearing of progeny.11 This circumstance has profoundly affected the part of women in economic activity. Most of the ways thought proper for an adult woman to have contact with the people beyond her family circle are in a married status. Few women escape marriage. Indeed, it is so normal an event that to pass the age of 25 unwed is considered a serious misfortune. To covet matrimony in vain does not carry with it any of the other possibilities (such as occupational careers) that exist in Western societies, and becomes a matter of anxiety if not humiliation for an entire family. All this does not mean that women under these conditions do not work. The lot of a "home-maker" is a more laborious one in agrarian than in industrial countries, and may encompass more different sorts of particular activities from case to case. A wide range of tasks, in­ cluding production of many physical commodities, is conducted as 10 If it were, it would cease to present any problem, for it could be reduced simply to bureaucratic patterns of large-scale organizations. 11 This does not mean that the particular family institutions are everywhere the same: such societies have long been renowned for the variety in such matters as mar­ riage ceremonies, types of education of children, adolescent behavior and the division of labor in specific household tasks. We are not here concerned with superficial varia­ tion, but with the general features which are shared.

THE DISPOSITION OF MANPOWER

a matter of course within the households that finally make use o£ them—items which in industrial societies are ordinarily purchased in the market. In these activities women are occupied often from childhood to old age, and sometimes more vigorously than men. It is this economy of the household, and not real inactivity, that uti­ lizes the efforts of females in useful pursuits but provides against their holding extra-familial occupations. We can be certain that a goodly portion of the work done in Tai­ wan was concealed in the activities of women. When such is the case, estimates of manpower based on occupational data are necessarily distorted, especially when these data have been compiled without overt care to this problem. Here the only feasible procedure in meas­ uring manpower is to exclude altogether this important segment— womanpower—and to concentrate only on the "male labor force" as representing the dominant working patterns of the population. Even within the male population, however, not all of its members are eligible or able for permanent employment. Close to half of all males in Taiwan—practically all those without occupations—were consistently of ages in which employment was not customary. Age always has a great deal to do with whether a person follows an occu­ pation or not, and often with the sort of work he may perform. Tasks that are looked upon as useful are commonly ones that infants and the aged, for example, lack the capacity for doing. In between these extremes, some males of certain ages are excluded from the labor force when sufficient value is attached to other activities such as school attendance, acquisition of special skills, patriarchy, or in­ firmity—either due to priority of the competing activities themselves (regardless of who performs them) or due to the kinds of persons involved. These various factors determining employment are so closely asso­ ciated with age that at any one time almost all of the men of young and middle adulthood in a population have some occupation. The exclusions are concentrated in the years of childhood and old age. For this reason it is meaningful to speak of the "population of work­ ing ages," and often more information is available about them as a whole than about the labor force itself. The Labor Force in Taiwan

Let us continue this line of inquiry by inspecting the portions of the population that are employed in relation to sex and age. With

THE DISPOSITION OF MANPOWER

some idea by this time of what the census standards of employment purported to be in Taiwan, one should then see how they are per­ formed in practice. As a first step it is interesting to compare the occupied portion of the population with a corresponding group else­ where. To judge from the proportions of men occupied by age in various countries—and under varied kinds of census administration— the extent of employment is approximately the same among the re­ corded populations of the world.12 In 1930, for example, these proportions in Taiwan closely re­ sembled in most ages those for the United States, when a similar standard ("gainful occupation") was used in this country to assign status in the labor force. In Table 17, the figures for these two coun­ tries were nearly alike among all men above age 25, though they were markedly higher in Taiwan at younger ages. The parallel does not imply that "work" amounted to the same thing among men in both populations. Nor does their divergent use of young men suggest that Americans were necessarily more prone to idleness than Taiwanese. The chief significance of Table 17 comes rather from showing that work as defined for census purposes commands the energies of prac­ tically all of the young adult men in both populations,13 despite the wide differences in economic organization between the two countries. Granted the basic similarity, the principal difference in Table 17 deserves comment. In the first place men in Taiwan evidently begin to do regular work much earlier in life than in the Western world. Extension of general education into the years of later adolescence, as everyone is aware, is a recent and modern phenomenon and con­ fined to a few countries, for it has just recently been felt to be indis­ pensable to the skills that this new world requires. The luxury of being able to maintain these services is not readily appreciated with­ out contrasting it with less fortunate areas. In Taiwan, as in most of Asia, practically everyone destined to have an occupation was al­ ready employed before age 20. Many, not shown here, even entered a permanent occupation, in which they remained, before reaching age 15." 12 Notestein, Frank W., et al. (1944), p. 118. is It is interesting to note that the differences made by different criteria—i.e., more refined labor force measures—are slight, if not trivial, for our purposes. Cf. Durand, John D. (1948), Appendix A, pp. 197-200. i* The census data showing dependency of childhood up to 1930 do not tell the whole story. In the latter 1930's the Japanese made concessions toward the extension of school­ ing, until by 1940 Taiwan was perhaps better in this regard than most Asiatic countries

THE DISPOSITION OF MANPOWER TABLE 17 THE LABOR FORCE IN THE TOTAL POPULATION, TAIWAN AND THE UNITED STATES

Proportions of All Persons in Each Age Group Who Had an Occupation, Male and Female, 1930 Age Groups

Taiwan (Taiwanese Only") United States (All Groups) All Occupied outside Occupations Agriculture All Occupations

Males

15-19 20-24 25-34 35-44 45-54 55-64 65+

85.4 96.4 97.5 97.3 93.8 81.8 54.7

24.4 33.6 35.9 33.9 29.0 21.6 11.9

47.8 89.9 97.3 97.6 96.5 90.2 58.3

5.6 4.6 5.2 5.7 4.4 3.0 1.3

26.6 42.4 27.8 22.6 20.4 16.1 8.0

Females

15-19 20-24 25-34 35-44 45-54 55-64 65+

38.0 32.0 31.2 31.2 23.3 13.4 4.3

* Domestic servants among the Taiwanese, male and female, have been excluded. Sources: U. S. Dept. of Commerce, Bureau of the Census, 15th Census of the United States, 1930, Population, Volume V, General Report on Occupa­ tions. Chap. 4, Table 2. Taiwan—Census of 1930, Prefectural "Volumes, Table 15.

For reasons advanced above, data on occupations for women are less satisfactory. In Table 17, roughly one-third of the women in most ages up to 45 reported some occupation in 1930; but practically all stated that they were "farmers."15 While superficially these figures appear to make a good showing against the data for the United States, the resemblance is misleading. In fact, on the basis of this information besides Japan itself. But these events occurred too late to have much effect on the ages of starting work. 1S Even for the few women who acknowledged occupations outside of agriculture, the specific kinds of work that occupied the bulk of this group are, suspiciously, the very ones that are usually conducted in households. For example, most of the "industrially employed" women were engaged in handicraft production (wood and bamboo articles, food processing, Panama hats) typically conducted on a household scale.

THE DISPOSITION OF MANPOWER

one can really say next to nothing about the employment of Tai­ wanese women. Henceforth, except for special purposes, we shall confine ourselves to using occupational data for men only. Occupational Composition of the Male Labor Force

With the outline of total employment among males of different ages now established, it is only a matter of course to follow up by showing the occupational composition of the same group. This amounts to a close view of the data in Chapter III, considered for the moment apart from the differences in age structure and differ­ ences in the extent of employment in various ages.

OCCUPATIONS, TAIWANESE MALES

1 5 - 3 4 YEARS OF AGE 19051 1920

j

1 9 3 0I 35-64

6 5 AND OVER 1905 1920 1930

20

40

60

80

100

PER CENT I AGRICULTURE I FISHING,MANUF., MINING I:-:-;1COMMERCE

TRANSP- 8 COMMUNICATION GOV'T,

a

PROFESSIONS

OTHER

OFFice OF POPULATION RESEARCH, PRINCETON UNIVERSITY

Fig. 12. Occupational Composition of Taiwanese, by Age, 1905-1930.

Even taken by separate age groups, Taiwanese men were perform­ ing the same types of work after 25 years of colonial rule as at the beginning. The situation can be efficiently summarized in Fig. 12, which bears out in some detail the evidence of the last chapter. As regards the whole of the male working population, the stability existing from one time to another was not exaggerated in Chapter III.

THE DISPOSITION OF MANPOWER

Broad features like these were reduplicated in a degree that over­ shadows all accompanying trends—in this case down to the disposi­ tion of manpower at all working ages at successive dates. Fig. 12 shows clearly that this society had, up to 1930, put new generations to work along lines very similar to those of the old. That these proportions should remain almost constant through time implies great stability of the agrarian social order. But this evi­ dence is of a nature that best illustrates the more prominent relation­ ships, and does not give much weight to the shifts that did occur. Farming activities again appear predominant—as in fact they were— to the point of eclipsing the trends in growth of other types of em­ ployment. These can be brought out only through the use of more responsive measures. Moreover, even this degree of stability does not mean that indi­ viduals could not have moved from one occupation to another. A circulatory movement of individuals is fully consistent with the data of Fig. 12, provided that the avenues remained unchanging. It is not only possible but implied by the data themselves. In Fig. 12 the swelling and contracting of employment among different occupations from lower to higher age strata of course represent only the positions of people in these ages at one time; but a pattern that is repeated so closely in three successive censuses can hardly reflect just the events of a short period. Over such a long span of time, unless the various occupations lost their members through widely different rates of death and retirement, this pattern could come about only from losing them through movement to other occupations. If part of this movement is a matter of such circulation, it demon­ strates all the more strongly the part of traditional institutions in maintaining the society against the forces of change. Movement of this type will not be readily distinguished from shifts as a result of actual change, when the two are later analyzed apart from the com­ position of the labor force. It is therefore important to make clear at this point that whatever the extent of such shifts (as yet undis­ closed by these techniques), there must always have been a prominent element of circulation involved. Movement between Occupations

The data of Fig. 12 are poorly adapted to illustrate the nature of occupational movements in the population. Agricultural employ­ ment is preponderant, and masks the shifts that individuals actually

THE DISPOSITION OF MANPOWER

made between the classes of occupations that are listed. All move­ ments in other lines of work hence appear small because they were absorbed in the numbers of all occupied persons, instead of being compared with the numbers that had been in each line before. Some moves of occupation always occur in conformity with tradi­ tional usages—like entry into the labor force itself, moving from one position to another with gradual acquisition of skill, the advancement of successful men and the decline of failures, and eventual retire­ ment from active work. Shifts such as these are found even among the peoples who are the most bound by custom, and do not depend on the existence of new social conditions. The problem in this sec­ tion is to develop a more sensitive procedure to detect some of these transfers. A more direct approach to this "hidden" movement is to locate persons in their later occupational careers after they once have started. Comparing the population in the same interval of age at two suc­ cessive points of time meant that we were dealing with different people each time. Their selection of occupations might have been the same at both dates, since one of the functions of tradition is to perpetuate a given system of choice in new generations; but this type of comparison might not reveal what happened to any of them after the first observation was made. Uncovering this subsequent be­ havior rests on finding the same persons at a later date, which means also at different ages. Age, as regarded by demographers, is one of the few characteristics carried by people that are predictable indefinitely into the future. Those who were in age 20 in 1920 will be exactly age 30 at the same time of the year in 1930. Accordingly, those persons who reported an occupation in 1920, and survived, can be located in an age interval ex­ actly ten years higher in 1930. By the simple device of assembling the appropriate pair of figures and removing those who died during the interval, occupied persons in any age group can easily be compared with their survivors at the later date. One can then find whether, as a group, they still remained occupied, and to what extent. The same basic reasoning applies as between one type of employ­ ment and another; the survivors of a group of farmers at one time, for example, can be compared with the number actually found at another. The difference in this instance would represent the net gain or loss to these farmers in relation to the remainder of the economy,

THE DISPOSITION OF MANPOWER

including entries, retirements, and transfers to other classes of oc­ cupations. As in all formal problems in which one seeks to determine the in­ fluence of some factors while removing others, this procedure re­ quires that certain conditions be met. First of all, the population should be one in which some significance may be attached to "net movement" between occupations. In an industrial society, occupa­ tional transfers are frequent: over a period of several years an indi­ vidual may make multiple moves, or certain occupations may be sub­ ject to many entries and exits among members of a given age group. If widespread, either of these types of event would make an estimate of net change meaningless, because they would distort it as an indi­ cator of the true number of moves. In Taiwan, with foreknowledge that few people changed occupations after first adopting one, it is safe to assume that net movement is a good approximation to total movement across broad classes of occupations. Second, it should almost go without saying that occupational cate­ gories used in this comparison should be comparable, if not identical, at the two points of time. This condition is moderately well satisfied among the different occupations as they are arranged here.16 Next, the risks of death at each age should, for perfect accuracy, be equal from one occupational group to another. It is unlikely that this condition could be strictly fulfilled in any population. In Tai­ wan, many of the pursuits that were conducted in cities, for example, were probably more healthful than farming, for the hazards of city life were not greater than those in the countryside. Here, the mortal­ ity estimates that are used must be derived from the average experi­ ence among all Taiwanese males. Since no allowance can be made for variations between different occupations, some unknown biases probably enter the computations from this source. Age reporting must be very accurate in order to give assurance that the persons being compared are the same. By defining age in terms of the date of birth, the censuses of Taiwan developed age re­ porting that was more accurate and consistent than in most Western countries. While between 1920 and 1930 there was a slight change in 16 With one exception, which will be pointed out below. There must also be reason to believe that the activities comprising a given occupation have themselves not greatly changed over the period. This is one condition over which the investigator has no con­ trol whatever. We can simply assume for Taiwan that it is met reasonably well for the period in question.

THE DISPOSITION OF MANPOWER

the technique of assigning age, it was systematic through all ages and cannot create bias either by age or by occupation. Finally, because mortality estimates are usually applied to a popu­ lation in the form of survival ratios, persons should be distributed evenly within the intervals of age that are used. When these intervals are broad, there is greater chance of introducing some inaccuracy due to uneven grouping within them, since mortality is known to vary with age. The intervals used for this manipulation all spanned 5 years of age, and some element of control is therefore lacking. This is especially true for occupied persons aged 15 to 19, the only group that departed greatly in internal composition from the general popu­ lation.17 There is no reason to believe that the imprecision arising from this source should create any appreciable bias by occupation. Even where certain occupations may have distinctive age patterns of recruitment or retirement that give them age groups with unusual internal structure (e.g., farmers aged 15 to 19), the bias is very small.18 These conditions would have to be met in order to apply the pro­ cedure we have outlined to make estimates that are precise in num­ ber. But the procedure itself is only an approximation. Nonfulfill­ ment of any one of the conditions would not necessarily be serious, and could be foreseen and minimized by some arbitrary correction. No allowance can be made for the interaction of all of them, and for this reason some caution is appropriate in interpreting the follow­ ing results. With these qualifications, the way is clear to construct a set of comparisons between 1920 and 1930 according to the outline above. We shall estimate the balance of men entering and leaving ("retir­ ing" from) the working force by subtracting the survivors of those who were already at work in 1920 from those who were found with employment in 1930. (The survivors from 1920 are advanced ten years of age: all ages given in the following tables represent ages achieved by 1930.) In the same way, we shall estimate the balance of 17 The mortality estimates were formed by life table ratios of LI+10/LX. For con­ venience, and because it made no perceptible difference in the results, values were used for only the middle year of age in each 5-year interval. The values employed were taken from tables for actual cohorts, developed in connection with a later part of this study. They thus represent the actual mortality experience of the middle age cohort in each 5-year interval. 18 Actually, to avoid bias by occupation from this source, the condition may be re­ laxed from that of even internal age distributions that are the same among occupations within a given five-year interval. This problem is therefore less serious than it would first appear to be.

THE DISPOSITION OF MANPOWER

movement into or exits from a given occupational class by subtract­ ing the survivors of those at work in that occupation in 1920 from those found at work in that occupation in 1930. The results of these operations can be designated as "gain" or "loss"; the ratios of these numbers to the calculated survivors as "per cent gained" or "per cent lost."19 Drawing upon the range of ages from 15 to 64 carries the com­ parison through the span of greatest employment. The figures furnish in cross-section a picture of occupational movement in any selected categories with respect to all the remainder. While a cross-section, it is not an instantaneous view. Each entry in the table is based on the experience over a 10-year period for the given age group. Hence the number of age groups that can be usefully included is restricted by the length of this period. Older men have been dropped at age 70, which means that this last group was aged 55-59 in 1920. Similarly, the youngest age group is that of men 25-29, since in 1920 they were only 15-19, and too few of the boys under age 15 claimed occupations in 1920 for them to be of great interest.20 The differences between survivors and actual workers in 1930, and the ratios in which they stand to the numbers of survivors, are the most useful forms for interpreting these estimates. In Fig. 13 these are shown graphically in the arbitrary categories of "gain" or "loss," as explained above. The bar diagram in Fig. 13 shows the actual numbers of net entries and withdrawals, during the decade 1920-1930, of all occupations taken together. It is of some special interest in establishing the shape of these movements at different ages throughout the male labor force, and as a standard against which to compare the patterns for separate occupational groups. More important here is the second part of Fig. 13, giving the per­ centages of the survivors of 1920 workers who had changed their occupational status between 1920 and 1930. Along the middle seg­ ment of the age scale, these percentages are practically zero (ages 30-44, covering the actual age span of 20-44). The working force as a whole was almost stationary over this entire interval, consisting of those persons who were already employed, with few entrants and few withdrawals in relation to the numbers who were already there. Over this span it is not likely that many entries and exits occurred is Where entries are still taking place, these are not true proportions: the numerator is not included in the denominator. 2o By a special arrangement, it will be possible somewhat later to deal more ade­ quately with the males 15-ig and to include the group 10-14.

THE DISPOSITION OF MANPOWER OCCUPATIONAL MOVEMENTSiTAIWANESE MALES, 1920*1930

PERSONS, THOUSANOS

PERSONS, THOUSANOS

20i

20 TOTAL

OUTSIDE

IN AGRICULTURE

AGRICULTURE .

IO

-10

20

'

'

t

40

'

I

t

60

40

20

i

20

60

l

40

i

l

60

•10

AGE

60

40

\

'V—OUTSIDE AGRICULTURE

20

0

T -20

IN AGRICULTURE

-40 -60

25

•"TOTAL

30

35

40

45

50

j

ι

L

55

60

65

70

AGE

OFFICE OF POPULATION RESEARCH, PRINCETON UNIVERSITY

Fig. 13. Occupational Movement of Taiwanese Males, 1920-1930. Net "Gain" or "Loss" Inside and Outside of Agriculture, and in the Total Labor Force.

so as to balance each other; clearly few men began or left off work completely (except through death) in these ages, and what move­ ment there was must have been limited to transfers from one occu­ pation to another. It was established above (Table 17) that the maximum degree of employment was reached by age 30. It follows from these data that in this decade almost everyone who was likely in his life-time to acquire an occupation had already done so at age 20.21 By similar 21 A small part of this pattern may possibly be a peculiarity of the unusual decade involved. The census of 1930 was taken during a world depression, which had a brief but severe effect in Taiwan. There are hints that this was reflected slightly in the pro-

THE DISPOSITION OF MANPOWER

reasoning, 90 per cent of all men who were going to become em­ ployed by 1930 must have held occupations already between ages 15-19 in 1920: since only 10 per cent of the employed survivors to this group had begun their employment later than 1920, though by 1930, at ages 25-29, the ones with occupations already accounted for almost all the population in that age. One takes for granted that the boundary lines between being oc­ cupied or not were somewhat indistinct among young men entering and old men leaving the labor force. Even so, Fig. 13 gives a picture of the age incidence of retirement just as interesting as its evidence for the stability of employment through earlier adult ages. Starting with the men who reached ages 45-49 in 1930, the rate of exits from occupations over the preceding 10 years rose sharply with advancing age. Of the Taiwanese aged 55-59 in 1920, more than half the sur­ vivors had retired before age 70. Earlier ages exhibit what appears to be a gradual progression toward this rate of retirement. This would hold true even if we assume that the world depression in 1930 forced older men out of the labor force more rapidly than in normal times. The pattern would remain the same; only its age spacing would change. These findings tend to conform to what the earlier chapters would lead one to expect. They do not provide much new information (except by way of verification of the procedure that is being used). Of greater significance are the patterns that hold among the occu­ pations that make up the labor force. This is shown in Fig. 13 where the bars also represent the numbers of Taiwanese males "gained" or "lost" in 1930 as compared with 1920. Agriculture is the most striking of these. There was an actual loss of several thousand young adult men who had been farmers between the ages of 15 and 30 in 1920. This does not imply that agriculture was failing to recruit new farmers. The losses in part reflect the fact that males undertook occupations at younger ages (at least in their census returns) in farming than in other pursuits. Agriculture gained many of its entrants before age 25 (the youngest age shown in these charts); by age 25 it may or may not have continued to reportions of older men claiming some occupation in 1930. Despite the tendency of the "customary occupation" census concept to resist inroads of short-term unemployment, these proportions are slightly lower in advanced ages than those in previous censuses. Even so, the "true" figure might make the proposition above even stronger, by re­ vealing "normal" retirement to have been slower than Fig. 13 suggests.

THE DISPOSITION OF MANPOWER

cruit new workers—something that this procedure can never reveal— but was in any case exporting enough young men to the rest of the economy to outweigh these and create a net loss of workers. Though substantial in the total number of workers involved in net movements among other occupations,22 this rural exodus fades into insignificance when related to the number of men in these lower ages who remained in agriculture (Fig. 12). Obviously the standard against which to measure these events has much to do with assessing their importance. Losses of young men from agriculture supplied a fair share of all the entrants to other occupations in these ages. But this movement was not enough to make more than a slight inroad in the store of manpower for farming—and was more than offset by recruitments with agriculture between the ages 10-14 and 20-24. In this age interval, agriculture absorbed more than half of the total entrants to all occupations from 1920 to 1930. If among the survivors of, let us say, all males 15-19 agriculture had merely matched its own recruitment to the scale of the total labor force (i.e., that of Fig. 12, above), two-thirds of the total en­ trants to the labor force in this age would have become farmers instead of entering other occupations as in fact they did. No less important is the stream of older men—40 to 54 in 1930— who were entering agriculture. Though numerically not a large group, they reveal a degree of movement at least as great as that of the persons who abandoned farming at younger ages.23 These were the ages in which "retirements" from other occupations were beginning. If we assume that the exodus of young farmers was a "normal" oc­ currence (and some of the earlier census data suggest this is so), it is almost certain that these older recruits to agriculture were re­ entrants or "returnees" who had previously left. The figures again reveal only net movement; if some farmers of these ages were "retiring" from agriculture, then returnees must have been more numerous than the actual figures here indicate. Evidently agriculture not only supplied young men to other parts of the economy, but reclaimed older workers that other occupations 22 It needs to be emphasized frequently that the procedure we are following does not give an exact estimate of the total number of men who left agriculture for other work, even in the decade in question. Some of the actual exits were probably balanced by new recruits, though the number at these ages was undoubtedly smaller for agriculture than any other occupation. 23 I.e., in view of the fact that not many men were alive in these ages, these entrants bear a larger proportion to the total number of farmers of the same age. Their sig­ nificance in this respect can be read from Fig. 13.

THE DISPOSITION OF MANPOWER

had discarded. Re-entry to agriculture was thus for some men an alternative to ceasing work altogether. The data for other occupational groups can best be examined in Fig. 14. The bar diagrams show the amounts of net gain or loss, which should be compared with the total amount in Fig. 13. The chart of percentages gain or loss gives the corresponding view of movement for each, taking into account at each age the number of persons who were already there. The first indicates the numbers of people involved; the second abstracts from this. Due to changes in census practice, certain of the chief occupational classes must be lumped together so as to be comparable between 1920 and 1930. These consolidations make it impossible to say very much about activities in the non-agricultural portion of the economy that is not already implied by the relation of agricultural employ­ ment to all of them. This sector can be split into three main parts to compare their uses of manpower: industry and commerce (also including fishing and mining), professional positions and public administration, and transport and communications. Though not of the same size, all three of these classes absorbed and released workers in a similar age pattern, and all were distinct from agriculture and from the total labor force (see Fig. 12 and Fig. 13). All were most conspicuously users of youth, and taken together they obviously de­ pended on people already working in agriculture for recruits. To some extent all, but especially transport and communications, tended to cast off older workers sooner than agriculture. Evidently workers in transportation and in professional fields or the administrative bureaucracy were recruited more exclusively while very young. This is not surprising, in view of the fact that most Taiwanese held only junior positions in these fields, and fre­ quently positions where some specific technical training put youth at a premium (for example, drivers of motor vehicles, railroad train­ men, school teachers, telegraph workers, clerks).24 Finally, a few words are required about the group of "other occu­ pations." The steady "loss" of persons from this category is a fixed feature of Taiwan's whole succession of occupational censuses. Under the assumptions applied in the rest of the occupations, Fig. 14 would appear to indicate a flood of departures, in all ages, from these mis24 However, transportation workers also included a number of pullers of rickshas and other carts, who would likewise tend to enter this employment while young, but for different reasons.

THE DISPOSITION OF MANPOWER

OCCUPATIONAL MOVEMENTS, TAIWANESE MALES, 1920 -1930 PERSONS, THOUSANDS

PERSONS, THOUSANDS

MA ISiUFACTURI WG,

TRANSPORTATION

COMMERCE, ETC.

COMMUNICATION

GOVERNMENT,

IO

PROFESSIONAL

20

40

60

OTHER

120

TRANSR, COMMUNICATION

80

40

GOVT., PROFESSIONAL

\

X OTHER MANUF., COMMERCE, ETC.

-80

25

30

35

40

45

55

60

65

70

AGE

O F F I C E O F P O P U L A T I O N R E S E A R C H , PRINCETON U N I V E R S I T Y

Fig. 14. Occupational Movement of Taiwanese Males, 1920-1930. Net "Gain" or "Loss" in the Occupations Outside of Agriculture.

cellaneous activities. More reasonably, it implies that most of these workers were shifted to more specific occupations merely by refine­ ments of classification in 1930. Accordingly, these reclassified per-

THE DISPOSITION OF MANPOWER

sons must have "entered" the remaining occupations. The data for "other occupations" are by themselves useless. Here they serve as an additional reminder that the estimates of occupational move­ ment in this section are inexact, and deserve to be treated with caution.25 Using only the figures for males who held occupations, it has not been worth while to carry out these procedures with those who fell below age 25 in 1930 (age 15 in 1920). It is formally possible to estimate the survivors after a decade to employed boys aged 10-14 in 1920, probably to a fine degree of accuracy. But so few had acquired occupations by that age, outside of agriculture, that the results would have been meaningless as a basis for deducing any kinds of movements. Yet this neglected age span—between age 10 and age 24—is highly significant. It is here that most males initially entered the labor force. At some point between these limits most Taiwanese under­ took their first regular adult employment, exercising whatever choice was available to them. The nature of a person's work for many years in the future tended to be fixed by this initial occupa­ tion. This stage in the occupational careers of most males has a special import in an agrarian society. Most of the attitudes and skills pertinent to livelihood are imparted during childhood in the household—occupations, in effect, tend to be inherited. Knowing the sorts of households where boys spend their childhood, it should be a test of whether they were adopted in accordance with tradition if we are able to predict their subsequent occupations. Taiwan's population statistics are not without resources perti­ nent to this question. In the census of 1920, following the practice of previous years, the census authorities did more than classify by their type of work just the persons who claimed an occupation. Each census schedule included a space for everyone not reporting an occupation to be entered as "dependent" or "without occupation." The "dependents" made up practically all of this residual group. They were taken to be those who performed minor household work, 25 One can almost rule out the possibility that errors from this source have spuriously created the patterns derived from the data. In the first place, the men who have been reclassified from "other occupations" were not concentrated in any special age range, but rather spread more or less evenly over the whole span of working ages. Secondly, it is most likely that these persons were re-allocated to the non-agricultural sector of the labor force, and not within farming itself. They could in that case not affect the most important comparison of all—the differences between agriculture and the rest of the economy.

THE DISPOSITION OF MANPOWER

excluding ordinary domestic servants. Tables were published as­ signing all persons in this group to the occupation on which they depended—in most cases, that of the household head. One such table, with this material cross-tabulated with age, can be used to supplement the data for men with full-fledged occupa­ tions. Here the "dependent" category amounts to a table of almost all the population outside those actually reporting occupations, classified by the type of work that they performed in the household. Virtually all of them (99.6 per cent) were members of the family of the household head. In other words, for young boys this table rep­ resents a division roughly along the lines of economic activity that supported their families. All except a small fraction of Taiwanese males attained some oc­ cupation sooner or later during adulthood. Most of those who were recorded as dependents at the pre-adult ages of 10-14 an^ 15"19 would soon pass into some regular occupational group, and so were not far from such a transition at the time they were enumerated. This peculiarity makes it possible to treat the data as an approximate distribution of pre-labor force males according to the source of live­ lihood of their families. From this information we can count virtually all the males in these ages who would ever enter an occupation in the sense we ascribe to the term here, classified by the type of preoccupational work they performed. The same operations that were applied earlier in this section can be repeated with the male population thus tabulated (occupied and dependents combined), arranged by age. The results are plotted in Fig. 15 for the decade 1920-1930. No comparable tables of depend­ ents were published after 1920. Hence, the survivors to occupied males plus dependent males of 1920, in each age interval beginning this time with 10-14, have been compared simply with men who had an occupation in 1930. Fig. 15 supplements the findings of Figs. 13 and 14 by incorporat­ ing the group of males that reached ages 20-24 in 1930. It adds nothing new to the results at older ages. In fact, a small number of "dependents" are found among males even in adult age groups—a few men always remained in this status. In the "working ages," their proportion to all males was very slight—as low as two per cent. The survivors to workers and dependents taken together, therefore, were slightly more than the workers found in 1930. Fig. 15 is subject to

T H E DISPOSITION OF

MANPOWER

Fig. 15. Occupational Movement and "Inheritance" of Occupations by Taiwanese, 1920-1930. Net Movement between Occupations, and Movement into Occupations by Pre-Labor Force Males.

a small amount of exaggeration in the "loss" at all adult ages from this source. For the total male labor force, the ratios of gain and loss over the decade 1920-1930 to survivors from 1920—by age in 1930—are connected in Fig. 15 by one continuous line. These ratios take into 96

THE DISPOSITION OF MANPOWER

account the survivors of prospective as well as actual members of the labor force, starting with age 10 in 1920. This provision depicts a total occupational system that was self-sufficient with respect to the supply of young entrants during the decade 1920-1930. In the pres­ ent set of data, next to no one was recruited into employment from outside. Therefore, deviations from this self-sufficiency in any occu­ pational class implies some shifting—some failure to "inherit" occu­ pations—and should also reveal the direction of the movement. If the assumptions are correct, farming households apparently supplied a greater proportion of their males in these earlier ages to the rest of the labor force than did agricultural employment at later ages according to Fig. 13. Since boys in young ages were more numer­ ous than men in their twenties, the numbers involved were much larger. Industry and commerce (also including fishing and mining) must have absorbed most of the young males who left agricultural households for other work. But they seem to have supplied an even greater number of their young workers over this period from their own families. As in Fig. 13, the fields of transport and communications, and pro­ fessional occupations and public administration, stand out as bor­ rowers of youth from other families. Extending this test to younger ages only accentuates their debtor position in gaining manpower. This gain is evidence that these fields were expanding, but expand­ ing by a very distinctive mode of recruitment. As before, "other oc­ cupations" indicate merely that the classification was inconsistent through time. But since the effect of this shortcoming is less among the ages specially discussed here, it need be of less concern than before. Census data in this form have thus given a comprehensive picture of the principal flows of persons moving from one occupation to an­ other in this decade. Even though the permanent result of the move­ ment (in terms of changing the age-specific occupational structure of the population) was very small, these materials show that changes of occupation were actually more numerous and followed welldefined patterns. In particular, agriculture had a double role. It was the outstanding source of supply of young men in other occupations; and it was a pursuit to which many men returned when they were no longer youthful. From this we may draw the inference that some diversification of livelihood can be absorbed by an agrarian society with a minimum of lasting effect: rural agriculture "lends" workers

THE DISPOSITION OF MANPOWER

to the rest of the economy in their young adulthood, and takes them back after their years of greatest contribution are past. Manpower and Age Composition of the Population

Unfortunately for the conclusions of this chapter and the previous one, no population data on employment of any consequence have been published for Taiwan since the census of 1930. All that exist for the very interesting period leading up to World War II are the economic statistics reviewed in Chapter II (of quality far inferior to census figures), undisciplined travel accounts of a few foreigners, and the quasi-official reports of the Japanese themselves—half secre­ tive and half boastful. For this reason there is special value in observing the size of the population of working ages, from which the labor force is drawn, and its development as a component of the total population of Tai­ wanese. Requiring only detailed information by age and sex, this important measure can be derived from more recent census ma­ terials which were not published with elaborate occupational tables. It can be carried up to 1940 with exact figures, and extended beyond this date by the use of special assumptions. All of the foregoing tables in this chapter have been arranged so as to represent occupational behavior of Taiwanese within certain age strata. This was a device for excluding temporarily two closely related aspects of labor force study: the age structure of the popu­ lation at any one time, and trends in the size of different age strata. Insofar as these factors help to fix the supply of labor and the amount of dependency in a society, they exert a profound and subtle influence on economic activities. At a given moment, among male Taiwanese some form of occu­ pation was almost universal in most of the span between ages 15 and 60. Narrowing attention to the men between ages 15 and 64, as the "population of working ages"—to follow a common convention— will obviously include all ages of extensive employment. Not many boys younger than age 15 held occupations, and few people survived at all after age 65 (not all of these remaining occupied). Therefore, the group of men aged 15-64 also contained practically all working males. Viewed as representing their region in the world, the Taiwanese are not one of the "major" populations of Asia. Consequently, we are free to analyze their composition without introducing the idea

THE DISPOSITION OF MANPOWER

of vast numbers, and the global problems that these tend to raise. Compared with some of these more familiar examples, in Asia and TABLE 18 AGE COMPOSITION OF TAIWAN AND SELECTED OTHER COUNTRIES, CIRCA 1930

Both Sexes Percentages of Total Population in Each Age Interval U.S.S.R. (incl. Ukraine)

United States

England b- Wales

Sweden

Age

Taiwan s

Japan

British India

0-14 15-39 40-64 65+

41.0 39.0 17.5 2.5

36.6 38.3 20.3 4.8

39.9 41.1 16.8 2.2

37.2 41.3 17.5 4.0

29.4 41.2 24.0 5.4

23.8 40.4 28.4 7.4

22.1 41.7 27.0 9.2

Total

100.0

100.0

100.0

100.0

100.0

100.0

100.0

" Taiwanese only. Source: Institut International de Statistique, Apergu de la Demographie, 1929-1936. The Hague, 1939. Table 7. Statistical Summary of Taiwan, Table 58.

the West, Taiwan ranked with India, a land of proverbial "popula­ tion problems." Together these two areas illustrate an ironic his­ torical circumstance: those areas that for other reasons are most handicapped by poverty are also the ones with the greatest burdens of dependency, and with the smallest proportions of their people in working ages. The differences are outstanding in the ages of maturity, near the range between age 40 and age 65, when Western nations have been in a position to make far better use of the accumulated training and experience among their people (see Table 18). While next to nothing is directly known of its earlier demographic history, Taiwan was already in this situation when it entered its census era, and was placed still more at a disadvantage as time went on. Through time, the population of these various ages in Taiwan has not shifted so signally as it has grown. From 1905 to 1940, each of these age strata almost doubled in size. In the same period, the important span 15-64 underwent slow but steady changes in relation to the rest of the population, older and younger. Its share in the total was reduced, though its numbers grew by more than one-half. On the other hand, the proportion of aged (always small) did not greatly change, though their numbers doubled. But most signifi-

THE DISPOSITION OF MANPOWER

TAIWANESE POPULATION, BY AGE GROUPS

15-39

AGE 0-14

40-64 65 +

1905 1940 (ESTIMATED) 1950

I

0

I

ι

20

I

ι

ι

ι

40 60 PER CENT

•

'

80

I

*

100

'1905 1940 ( ESTIMATED ) 1 9 5 0

MILLION O F f I C E OF P O P U L A T I O N R E S E A R C H , P R I N C E T O N UNIVERSITY

Fig. 16. Taiwanese Population in Working and Dependent Ages, 1905, 1940 and Projected to 1950.

cantly, infants and children, represented by all persons under 15 years of age, expanded into an ever larger part of the total popula­ tion (see Fig. 16).26 This young and dependent segment did not ad­ vance its share in rapid strides; only by 1940 did it reach 44 per cent of all Taiwanese. A glance at Table 18 will show how extremely this departs from the composition of populations in industrial countries. There is little reason to suppose that the trend would have con­ tinued much further, even with the large numbers of births that contributed to it each year and the efforts at health control already sponsored by the Japanese. On the purely hypothetical assumption of continuance of mor­ tality and fertility through the 1940's as in the period 1936-1940, a bar has been added to Fig. 16 for 1950. It gives an indication of what these same circumstances might have produced in ten more years. As an actual estimate it is of no value, since conditions did not in fact remain so favorable after the end of Japanese rule. The result­ ing figures are nevertheless of interest, for they suggest that without new improvements in longevity this trend might have progressed already as far as it could go.27 26 Fig. 16 is based on data for males and females combined, since when it comes to dependency of childhood and old age both required support. It is implicitly assumed that women in prime adult ages are not dependent in this same sense. 27 It is interesting to note that the age composition of Taiwanese in 1940 was almost identical with that of the "stable population"—i.e., the age composition that would

THE DISPOSITION OF MANPOWER

With continuance of these demographic conditions—and infinite resources for their support—the Taiwanese could probably have expanded in this same scale forever. But, while the pattern of occu­ pational movement that we have observed was also a stable one that did not contain many elements of change, it could not have per­ petuated current conditions indefinitely, any more than the actual growth has been allowed to continue without setback. The phe­ nomenon of growth itself was an underlying factor of instability for the society. If the Japanese had remained in control, they would have been forced to recognize the problem sooner or later, and either provide additional opportunities for livelihood or risk disruption of their program from other sources. Their efforts toward indus­ trialization were too small and too selfish to have made any difference in the result. have resulted from indefinite persistence of the fertility and mortality rates of the same period, 1936-1940. The rate of increase implied by the structure of this hypothetical stable population is somewhat higher than that which was actually recorded at the time.

CHAPTER V

Migration and the Growth of Cities DEPENDING on the way one looks at it, the history of colonial rule in Taiwan can give the appearance either of having brought great transformations to the island or of having left the indigenous social structure virtually intact. This apparent contrast comes from shift­ ing attention between the two principal groups of people that were involved—from the Japanese, who had little to do with the archaic customs and usages of the population, to the Taiwanese, who had only small acquaintance with the modern changes that were being introduced. Here we shall be concerned with those of the latter who came into contact with the new elements. Though the full force of these developments passed them by as a group, the Taiwanese were not all uninfluenced. The organization of farming, to be sure, continued as before with little modification. The activities most affected were those outside of agriculture. Hence the people must have moved to the new conditions, since the inno­ vations did not move to them. The Territorial Frame of Migration

Taiwan is a small island, and recording the migration of its people should not be a task of great magnitude or difficulty. Japanese re­ garded such a record as essential in keeping a close watch over the population. It was so important that they inaugurated continuous population registers for every village, every group of households. These reports were compiled in fine detail by the authorities and published in a single volume each year.1 A certain amount of adap­ tation is needed to make these registers answer questions that are germane to our interests. Japanese conquest began to provide not only the first reliable data for assessing movement in the population, but also a new type of control over it. It put an end to the historic influx of people from China that had been responsible for populating most of the island. ι The data proved to be more efficient for their purpose than for ours. The system faithfully counted the arrival of new persons in a community (the most important in­ formation from the standpoint of regulation) but did not expunge the records of all persons who left. The result was a more or less consistent bias toward overestimating the number of people living in a given locality, and of course a somewhat distorted view of total migration.

MIGRATION AND THE GROWTH OF CITIES

On a much smaller scale, Japanese replaced Chinese as immigrants. Chinese arriving from the mainland were given a distinct legal status as aliens. Their activities were restricted, and they were no longer allowed to merge with permanent Taiwanese residents. Practically all movement by Taiwanese thenceforth consisted of internal migra­ tion, redistributing the population that was already there. Only a part of the island is level enough to have encouraged heavy agricultural settlement; almost half of the total area is more than 500 meters above sea level. For administration, and therefore for the census, Taiwan was separated from 1920 on into seven main di­ visions. Two of them, plus an eighth for P'enghu (the Pescadores Islands) off the West coast, were technically under a different sort of political organization; but for convenience we shall refer to all eight as Prefectures.2 Most types of census and registration statistics were tabulated for each of these political units as well as for the total island. Beneath the level of Prefectures, local administration was organ­ ized around a set of Districts and a smaller number of Municipali­ ties. In the population statistics, these formed two parallel units: no District contained a Municipality, and both were subordinate to Prefectures.3 There were seven Municipalities at the census date in 1930, raised to nine before 1935, and 53 Districts in both years; in comparisons of different years, all will be standardized to the ar­ rangement in 1935. Inside each District were towns and villages. Few materials are available for these individually; for most purposes the smallest divisions that can be studied are the Districts.4 Finally, the mountainous interior region—the Aborigine Terri2 If not literally accurate, the term is appropriate in that authority below the Gov­ ernment-General was highly centralized in the subordinate governors of these divisions. How they are designated here is a matter of no consequence, as it made little difference in the actual administration. P'enghu, scarcely known to Westerners except as the Pescadores Islands, was in 1926 made a separate division under the Government-Gen­ eral. Prior to this it had shifted between separate status and a subdivision of Kaohsiung Prefecture. Hence there is no unbroken series of data for it. 3 Again, the details of political organization were not uniform, but that need not concern us here. Both terms are arbitrary: "District" for the term chiin (in its Chinese reading), and "Municipality" for shih. It was the practice to form new municipalities by removing some important town of sufficient size as a unit from its surrounding Dis­ trict, sometimes adding some former villages, also intact. This practice gives a basis for identifying the population of cities for comparison over a period when such admin­ istrative changes were taking place. 4 The area of towns and villages exhausted the entire area of Districts, leaving no space between their boundaries. Therefore they do not always refer to meaningful units of settlement, and the terms are not consistent.

MIGRATION AND THE GROWTH OF CITIES

tory—was put under special control (though it nominally fell inside some part of each prefecture). This area must be considered as a separate division of the island. Its actual conditions and the quality of its population data are distinct from the remainder. Japanese were secretive about their treatment of the Aborigine Territory, and we have little information about it that is dependable. These civil divisions were the outcome of some experimentation in government administration. The scheme just described was put into essentially its final form in 1920. Accordingly, most compari­ sons involving internal boundaries cannot be carried back beyond that year. The Distribution of Total Population and of Migrants

In 1930, the population of Taiwanese was divided among the Pre­ fectures as indicated in Table 19. Data in this form do not readily TABLE 19 TOTAL POPULATION OF TAIWAN, 1930

By Prefecture, Inside and Outside the Aborigine Territory (All Ethnic Groups) Area of Regular Administration

Prefecture

Popu­ lation

Taipei 896,380 Hsinchu 641,479 Taichung 990,571 Tainan 1,155,287 Kaohsiung 601,547 Taitung 46,182 Hualien 72,555 P'enghu 60,124 Total Island

4,464,125

Density As % of Density: Average Persons For Total Per Sq. Km. Island

Aborigine Territory

Density: Persons Per Population Per Sq. Km. Cent

9.4 11.8 6.5 9.0 11.1 5.0 4.4

279.3 233.9 210.5 46.2 53.7 473.9

141 105 120 100 90 20 23 204

17,151 23,232 24,975 4,359 31,772 12,619 14,304 —

—

-

232.7

100

128,412

7.6

3

328.0 243.9

4 5 3 4 5 2 2

Source: Census of 1930, Statistical Tables, Table 11.

reveal relative concentration and dispersion of people. The real centers of population, if we disregard for a moment the dense settle­ ment in the largest cities, were in portions of Taichung and Tainan

MIGRATION AND THE GROWTH OF CITIES

Prefectures, and in Taipei. A thin belt of level land along the east coast was only freshly settled; people were extremely sparse in the interior region (see also Fig. 2). Separating the Aborigine Territory removes the greatest variation in terrain and land use. On this basis the more populous sections of the island can be roughly distin­ guished, though only in terms of Prefectures (which were not evenly settled throughout). To facilitate this comparison, the density of total population (per square kilometer) in each Prefecture is ex­ pressed in the final columns of Table 19 as a percentage of the density over the entire island with the Aborigine Territory excluded. One can get a quick glance at the total effects of migration by examining the disposition of people among the smallest civil divi­ sions at two separate dates. This information is prepared, for the years 1920 and 1935, in Table 20. The intervening period covers TABLE 20 TAIWAN. TOTAL POPULATION BY SIZE OF CIVIL DIVISION, 1920 AND 1935

Excluding the Aborigine Territory 1920 Size of Civil Division

Under 10,000 10,000-20,000 20,000-30,000 30,000-40,000 40,000-50,000 Over 50,000 Total Island Under 10,000 10,000-20,000 20,000-30,000 30,000-40,000 40,000-50,000 Over 50,000 Total Island

1935 Population

Civil Divisions

A. Numbers 765,772 116 1,980,510 141 409,149 19 166,543 5 47,921 1 239,342 2

1,244,749 2,262,691 606,831 119,785 469,151 384,973

130 118 22 3 7 2

3,609,237

5,088,180

282

Population

Civil Divisions

284

B. Per Cents 21.2 40.8 54.9 49.6 11.4 6.7 4.6 1.8 1.3 0.4 6.6 0.7 100.0

100.0

24.5 44.5 11.9 2.3 9.2 7.6 100.0

Sources: Census of 1920, Descriptive Report, Appendix, Table 8. Census of 1935, Statistical Tables, Table 3.

46.1 41.8 7.8 1.1

2.5 0.7 100.0

MIGRATION AND THE GROWTH OF CITIES

fifteen years o£ uninterrupted growth in the island's population. This is a short span, but it is the longest one for which appropriate figures exist. Before 1920 the territorial divisions were not the same, and for the census of 1940 no full reports have been published. Apparently, from Table 20, the fact that there were more people in 1935 made little difference in their mode of settlement. At both dates more than two-thirds of all people (here including Taiwanese and all other groups) were living in places of less than 20,000 in size. Some of the larger civil divisions shifted upward from one group to a higher one; this is to be expected from the mere fact of general population increase. What is surprising is that the biggest cities—and as late as 1940 Taiwan boasted only three larger than 100,000—failed to advance substantially over the rest of the island. Subsidiary centers of less than 50,000 population made the most headway in relation to other places. A certain amount of administra­ tive reorganization enters into these figures, since evidently a few of the minor units were split up into still smaller ones. But these changes were confined to units of the smallest size, and did not en­ compass many people. Viewed in entirety, migration therefore did not have a marked effect in redistributing the population growth that was in progress among the Taiwanese.5 Economic activities had changed so little, however, that migra­ tory movements could hardly have been large. Besides, no large areas of new lands were available to farmers after 1900. Most of the additional space brought under cultivation since that time was con­ verted to agricultural use within the settled area. Expansion of agri­ culture along these lines did not open opportunities of wider settle­ ment.6 In the absence of empty lands, there was little for peasant farmers to gain in changing residence unless they also changed their livelihood. In general, if they remained agriculturists it was advan­ tageous also to remain in localities where they were well known, where their families were well established, and where they did not suffer the invidious treatment commonly accorded to strangers. 5 Difterences in natural increase, which might have hidden the effects of some migra­ tion in these figures, bore almost no relation to the differences in actual population growth between civil divisions. β The exception to this rule was the rugged east coast of Taiwan. There the authori­ ties brought in "colonists," first Japanese and later Taiwanese, to establish special settle­ ments in the limited plains area for cultivation of sugar and other crops. Though mak­ ing up a large part of the population later found in this region, this migration was small and of a very special nature.

MIGRATION AND THE GROWTH OF CITIES

On the basis of the previous chapters we should know in advance some of the reasons why people moved. If movement was difficult and uncommon within agriculture, then permanent migration was presumably a matter of going to new jobs. Occupational movement and spatial movement were two interconnected parts of the same process. This meant, for the society, a shift in its economic base; for individuals, it amounted to a change in status, together with exit from rural life and activities. Bearing in mind that the migrant population was small, we can try to discover where these people went and who they were. For the aggregate of those who crossed local boundaries, both origins and destinations can be found by estimating the balance of persons who moved into or out of each civil division. This measure is suited to detecting various types of migration. In its results much depends on the types of areas and the period of time chosen for making the estimates. Districts and Municipalities are the places that will be used here; they are small enough not to hide very many moves of more than ten or twenty miles. The small­ est time interval is set by the fact that census data give the best indi­ cation of total population change by place (i.e., the difference in the people counted from one census to the next in each unit of enumeration). This fixes the period to be used at five years. It is probably sufficient to cancel out some short-term migration and give a certain amount of weight to permanent shifts of residence. But intercensal population change represents the effects of both net migration and the net amounts gained by births and lost by deaths or natural increase. Since only a tiny fraction of births or deaths went unregistered, they give a good basis for taking out the component of natural increase. The residual term that remains from intercensal population change can therefore be regarded as "net in­ ternal migration" in any given civil division.7 From 1925 to 1930, net migration viewed in this fashion did not vary greatly among Districts. It is interesting, however, to see the regional groupings of Districts, in terms of per cents of their Taif That is, it is practically zero for the island as a whole. Discrepancies are possible for local areas if the chances of dying have been affected by whether or not some of their people were migrants, or if deaths were registered in a previous place of residence. There is no direct evidence available as to the error from those sources. Because Tai­ wan's population records were of high standard, and because migration was not heavy, it seems safe to assume that these distortions are minor in any particular place. The procedure used for estimating net migration may, however, increase the amplitude of errors in the residual values.

MIGRATION

AND

THE

GROWTH

OF

CITIES

Fig. 17. Net Migration of Taiwanese between Civil Divisions, 1925-1930.

108

MIGRATION AND THE GROWTH OF CITIES

wanese inhabitants received or lost in this way from 1925 to 1930 (see Fig. 17).8 These groupings divided the rural areas into two parts—a zone of out-migrant Districts in the north, and a more or less neutral region having almost no increase or loss through migra­ tion. A process of rapid settlement was beginning in the habitable parts of the eastern coast, where the authorities and a few private companies encouraged people from other Districts to go. This settle­ ment was carried on partly through government design, and was not entirely spontaneous. It grew rapidly from a small start, but did not shift large numbers of people. These transfers of people were extraordinary. The only other noteworthy exceptions to near-immobility over the rest of the island were cities. On the map these appear as the several isolated points with relatively high gain in migrants.9 They were interspersed in rural areas of migration deficit, suggesting that much of the move­ ment covered only the short distance from one civil division to the next. The same data used for Fig. 17 yield more information in terms of the relation of intercensal change with natural increase and net migration. Fig. 18 illustrates the pattern of association that holds in either case, taking each District and Municipality simply as a unit, equally weighted. Against total population change as the standard of comparison, the percentage of natural increase varied from Dis­ trict to District more or less at random, and within a narrow range. Net migration, on the other hand, contributed most of the variation in population change from one civil division to another. The gist of Fig. 18 is summarized in Table 21, this time with each civil divi­ sion weighted by the number of persons. In a somewhat different way, Table 21 reinforces what we have just learned about the com­ ponents of population change: natural increase was nearly the same in all quartile groupings of places according to total population growth. Only net migration varied consistently with total change. s Strictly speaking, ratios of estimated migrants to ig25 population. The Aborigine Territory has been excluded in these data, as it was only in the process of being "paci­ fied" at this time; the ratios therefore show the total area illustrated in Fig. 17 to have a net out-migration, presumably because some people moved into the fringes of the interior area as it was gradually subjugated. 9 In addition, part of one of the Districts was marked by exceptional growth among its neighbors due to the influence of a city that it contained. This was P'ingtung City, slightly to the east of Kaohsiung. It was elevated to a Municipality in 1932 and sep­ arated from the surrounding District.

MIGRATION AND THE GROWTH OF CITIES COMPONENTS OF INTERCENSAL GROWTH, TAIWANESE, 1925-1930 NATURAL INCREASE S GROWTH .

.

.

;«·

• ·· •

···

·

·\Χ\

_L '2

_L

8

_L

16

20

ι 24

28

_L

32

36

40

44

PER CENTINTERCENSAL GROWTH

28

NET MIGRATION & GROWTH

24 20

16 12 8 4 ·χ·

#

··

#

• .· * .·· · ·;

O -4

DISTRICT ·

.-

'*

MUNICIPALITY χ

- 8

I

-I 2

8

12

_L 16

20

J 24

L 28

J

32

L

36

40

44

PER CENT INTERCENSAL GROWTH OFFICE OF POPULATION RESEARCH. PRINCETON UMIVERSlTY'

Fig. 18. Net Migration and Natural Increase by Civil Division, Related to Intercensal Growth of Population, 1925-1930.

The residual estimates of net migration above indicate that city populations gained the most from migration. This fact does not of itself imply that they gained the largest numbers of migrants, for we have not yet established that the cities were large enough next to the rural population to have such weight. All that these data re­ veal about the migrants is, in an aggregate sense, the places of origin and destination. For some idea of the numbers of people involved,

MIGRATION AND THE GROWTH OF CITIES TABLE 21 NET INTERNAL MIGRATION OF TAIWANESE, 1925-1930

By Districts, Grouped in Quartiles According to Total Intercensal Change. With Municipalities Shown Separately* Per Cent lntercensal Change

Per Cent Natural Increase

Per Cent Net Migration

First Quartile Second Quartile Third Quartile Fourth Quartile

6.2 9.5 12.7 17.3

12.4 12.1 13.1 13.9

-6.2 -2.6 - .5 3.3

Municipalities

19.1

10.3

8.9

* Percentages are ratios based on population of 1925 in each civil division. Sources: Census of 1925, Statistical Tables, Table 1. Census of 1930, Statistical Tables, Tables 1 and 2. Vital Statistics, yearly volumes, 1926-1930.

and of the proportion of migrants who went to cities, we should rather deal directly with those who migrated. The census reports for 1930 contain a special table showing per­ sons born in a Prefecture other than the one in which they were enumerated—i.e., who had travelled across a prefectural boundary at some time after birth. This material is not in sufficient detail to establish the broad pattern of migration illustrated above,10 but forms a valuable supplement to our estimates. In particular, it sup­ plies a fuller answer to the question of relative drawing power be­ tween country and city. Very few Taiwanese shifted residence across prefectural bound­ aries—obviously fewer than those who moved across District bound­ aries. The ones who did are a specially selected sampling of all people who moved. Most changes of residence in an agrarian society may be expected to be over short distances. These involve the least break with the past and exposure to unfamiliar conditions. They contain much "normal" movement of people having nothing to do with social change, for they are often bound closely to stable local 10 It is, for example, an independent source of corroboration for the general spatial pattern of migration shown above. With it we can locate those migrants whose travels happened to cross prefectural boundaries (and who survived to the census date) ac­ cording to their place of destination. By this criterion, the places of highest immigration are mostly the same as those shown by the measure of "net migration" above in the text. These census data lack, however, any information about the District of origin of the migrant population, which is just as essential to know.

Ill

MIGRATION AND THE GROWTH OF CITIES

customs.11 Since the Prefecture in Taiwan was a fair-sized geograph­ ical unit, much of this short-distance movement regularly occurred within these boundaries and would never come to light in such coarse tabulations. The rural Districts receiving most migrants from other Prefectures were, not unnaturally, those which lay on the border of an adjoining Prefecture. The growth of each Prefecture through migration was therefore strongly conditioned by this fact. Cities were not so affected. Unless they were located near the edge of their Prefectures—and none was situated at a boundary—they had less chance of receiving shortdistance migration from another Prefecture than did their rural hinterland. Every Municipality in 1930, however, had a larger proportion of inter-prefectural migrants in its population than the remainder of its own Prefecture (see Table 22). These migrants, of course, were not necessarily people who had entered within a definite period of time, but those who had ever migrated. The east-coast Prefectures (Taitung and Hualien), which were non-urban, had the largest pro­ portion of people from other Prefectures; but we have already seen that they were exceptional cases.13 Significantly, the attraction of cities for people of other Prefectures tended to favor males more than that of their surrounding areas. In Hsinchu Prefecture (which on balance steadily lost migrants), the difference is greatest of all: the sex ratio of migrants to Hsinchu City was 167, to outside the city 72 (see Table 22). These facts are put into better focus in Table 23. When migration data of this type cover a complete area, they have the special virtue of enabling one to count the migrants by place of origin as well as place of destination. Part A in Table 23 more nearly corresponds to the residual approach that was adopted above. The column next to the number of in-migrants shows their proportion among all resi­ dents; the final column gives the proportion of all in-migrants who went to cities. At the same time when 23.5 per cent of all interprefectural migrants had gone to cities, the cities themselves held only 10 per cent of all Taiwanese. A more appropriate comparison than over the entire island is to examine the cities of each Prefecture 11 As, e.g., the "marriage migration" in certain countries. Cf. Davis, K. (1951), pp. 111-114. 12 And Hualien contained a growing city of moderate size, which was made a Mu­ nicipality in 1940.

MIGRATION AND THE GROWTH OF CITIES

TABLE 22 INTER-PREFECTURAL MIGRANTS IN TAIWAN, 1930

Taiwanese Residing in Prefecture Other Than That of Birth, Enumerated in Rural Districts and Cities B. Principal Cities (Municipalities in 1930)

A. Rural Districts Migrants from Another Prefecture (i.e., Born in Other Prefecture)

Prefecture

Total Taiwanese % of All ResiRest- Sex dents Number dents Ratio

Taipei

588,165 18,443

3.1

95

Hsinehu

611,145

1.1

72

Taichung

943,314 36,247

Tainan

989,591 21,678

6,989

Kaohsiung 552,243 24,985

Total Taiwanese % of All ResiResiSex dents Number dents Ratio

147,620 8,395 51,337 3,353 928 39,361

39,431 /76,073 2.2 100 147,615 45,288 4.5 121 3.8 111

Taitung

53,684

Hualien

72,103 25,670 35.6 136

P'enghu

55,936

Total Island

Migrants from Another Prefecture (i.e., Bom in Other Prefecture)

Name of City

5.7 124 Taipei 6.5 145 Chilung 2.4 167 Hsinchu

5,040 12.8 106 Taichung 5,101 6.7 110 Tainan 5,114 10.7 103 Chiayi 15,698 34.7 127 Kaohsiung

7,411 13.8 140

554

1.0

—

3,866,181 141,967

3.7

—

Source: Census of 1930, Prefectural Volumes, Table 30 in each.

separately, where their drawing power stands in clearer contrast to that of their surrounding rural area. Part B of Table 23 is an interesting supplement to the information already at hand. The deficit standing of Hsinchu is more conspicuous here than in Fig. 17, for the other Prefectures retained most of their own migrants inside their boundaries. P'enghu Islands, in addition, had lost relatively more people to other parts of Taiwan than any other Prefecture. These islands experienced disastrous storms and rural poverty for many years, and had no adequate economic base for many people to follow pursuits other than farming. Through the years P'enghu supplied an unbroken stream of people to Taiwan proper, most of them going to Kaohsiung City.

MIGRATION AND THE GROWTH OF CITIES TABLE 23 TOTAL SURVIVING MIGRANTS BETWEEN PREFECTURES, AND THOSE WHO WENT TO CITIES*

Taiwanese Enumerated in 1930 Whose Place of Residence Was Different from Place of Birth A. By Prefecture of Destination Resident in Prefecture, Born Elsewhere

Prefecture

Taipei Hsinchu Taichung Tainan Kaohsiung Taitung Hualien P'enghu

Number Who Migrated

% InMigrant of %of All Persons the InResident Migrant in Each Who Went Prefecture to Cities

30,191 7,917 41,287 31,893 40,683 7,415 25,650 544

3.9 1.2 4.2 2.9 6.8 13.8 35.6 1.0

38.9 11.7 12.2 32.0 38.6

Total Island 185,580

4.3

23.5

— — —

B. By Prefecture of Origin Born in Prefecture, Resident Elsewhere

Number Who Migrated

% OutMigrant cIoof of All Persons the OutMigrant Born in Each Who Went Prefecture to Cities

28,665 91,360 16,569 19,950 9,719 2,066 1,681 15,570

3.6 12.4 1.7 1.8 1.7 4.3 3.5 21.9

19.2 16.3 20.6 27.4 26.0 3.4 6.1 74.7

185,580

4.3

23.5

* People living in boats are excluded. Source: Census of 1930, Prefectural Volumes, Table 30 in each.

Movement to Cities: The Rate of Growth

These census materials above all emphasize one important fact: cities were the outstanding recipients of people from the rest of the island. The data regrettably can be used to demonstrate the case in only a limited way. The limitations are, however, of a nature that make this testimony an understatement—the total number of mi­ grants to cities was much larger than merely those who had crossed the borders of a Prefecture. Lacking further data directly for the people who moved, we may find out more about them at their des­ tinations as members of the urban population in Taiwan. Since these several cities took such a large share of all migrants, they must have grown more rapidly in population than the rest of the island. The Municipalities were only a part of all the cities and towns in Taiwan, for the distinction was one of administrative con­ venience. Not all had equally developed economic services. They are

MIGRATION AND THE GROWTH OF CITIES

a makeshift way of distinguishing urban places from rural Districts. Many subsidiary market towns, for example, must be overlooked for lack of data. But in a population that was predominantly rural, these cities were the most "urban," the farthest removed from traditional peasant agriculture. Under the watchfulness of the Japanese, a town could not enlarge its economic importance for the rest of the island very far before it was given regular municipal administration. Hence we may treat them as representing the most extreme departures from rural conditions to be found in Taiwan, and can proceed to find out how far they deviated from the rest. In 1932, two more cities were raised to municipal rank. The earlier and later population figures for all nine that were Municipalities in 1935 have been sorted out from the respective census reports, and appear in Table 24. It is clear that the cities of Taiwan were not great metropoleis. These nine were the largest, and many Westerners would regard only the foremost among them (Taipei) as being urban. One cannot gainsay their importance to Taiwan: they differed consistently enough from the rest of the island not to be lightly dis­ missed. One distinction that is obvious in Table 24 is in the rate of growth. Though the cities varied considerably among themselves, every one increased its Taiwanese inhabitants more than the entire island did. Most cities maintained steady rates of growth, and therefore tended to keep their differences with the others. The data are condensed and plotted on a logarithmic scale in Fig. 19, so as to contrast the rates of increase of city and non-city Tai­ wanese and Japanese. The inequalities in growth stand out clearly. Those cities that were expanding more slowly are likewise the ones that had relatively fewer migrants in the data above. They did not outgrow the non-city population by much of a margin, either in Taiwanese or Japanese. The differences in increase persisted over the entire recorded period, and carried over into the later years of ac­ celerated economic growth. Indeed, from these data on population redistribution one could scarcely guess that the tempo of economic life had changed after 1935. Whereas these same cities had held 10 per cent of all Taiwanese in 1920, they contained only 14.7 per cent as late as 1940. For all their growth, cities did not greatly surpass the increase of all Taiwanese.

MIGRATION AND THE GROWTH OF CITIES TABLE 24 POPULATION OF TAIWANESE AND JAPANESE IN NINE MAJOR CITIES AT SUCCESSIVE CENSUS DATES, 1915-1940* Batio: 1940/1920

1915

1920

1930

1940

98,495

108,107 33,692 28,951 22,250 32,810 62,115 31,392 27,153 18,420

147,620 51,337 39,361 39,431 40,747 76,073 47,615 45,288 28,917

221,602 69,578 51,060 54,741 122,173 80,828 96,588 46,961

2.0 2.1 1.8 2.9 1.7 2.0 2.6 3.6 2.6

364,890 3,466,507

516,389 4,313,681

808,415 5,510,259

2.2 1.6

45,211 11,898 2,649 8,817 786 12,140 5,445 8,840 3,112

67,687 18,541 5,102 13,214 1,977 14,955 8,575 15,437 4,814

91,550 24,815 7,504 16,300 2,977 16,654 9,960 28,336 6,747

2.0 2.1 2.8 1.8 3.8 1.4 1.8 3.2 2.2

98,898 164,266

150,302 228,281

204,843 312,386

2.1 1.9

TAIWANESE

Taipei Chilung Hsinchu Taichung Changhua Tainan Chiayi Kaohsiung P'ingtung

18,444 56,297

Total, All Cities Total Taiwanese

64,884

JAPANESE

Taipei Chilung Hsinchu Taichung Changhua Tainan Chiayi Kaohsiung P'ingtung Total, All Cities Total Japanese

37,974

6,817 9,708

* Cities not yet formed before 1935 have been reconstituted for previous years ac­ cording to their 1935 boundaries from the detailed census tables. Cities changed after 1935 have been adjusted to 1935 boundaries. Sources: 1915 Population Register ) (Both of these volumes contain the census fig1920 Population Register j ures for minor civil divisions.) Census of 1930, Statistical Tables, Tables 1 and 2. Census of 1940. (Figures secured from the Gazette of the Taiwan Govern­ ment-General, April 22, 1941.)

Japanese on the whole had stronger preferences for the same cities as did Taiwanese. But the functions of urban growth were not the same in the two groups. Taiwanese moved to cities as the most prom­ ising alternative to farming. Japanese, coming from outside to city and countryside alike, held technical and professional positions in

MIGRATION AND THE GROWTH OF CITIES POPULATION GROWTH IN CITIES

1,000

1,000

TOTAL CITIES 500

500

• OUTSIDE CITIES ( T I M E S 10)

200

200 TOTAL CITIES-^.-' -

OO * OUTSIDE CITIES 50

50 CITIES

A—TAIPEI 20

CHILUNG HS1NCHU CHANGHUA TAINAN

TAIWANESEJAPANESE—

IO 1920

_L

I

_L

1930

3--TAICHUNG CHIAYI KAOHSIUNG P'lNGTUNG

L 1940

20

IlO

1920

1930

1940

OPFlCE OF POPULATION RESEARCH, PRINCETON UNIVERSITY

Fig. 19. Growth of Population in Major Cities of Taiwan, 1920-1940. Taiwanese and Japanese.

both places. They certainly esteemed urban more than rustic life; but work of a similar nature was also available to them outside the cities. For them the place of residence had less influence on their type of livelihood. E t h n i c C o m p o s i t i o n of Cities

The growth of cities depended on only a small part of all Tai­ wanese to supply the necessary migrants. It was not so with the Jap­ anese. In a sense, the majority of them were migrants, having been born outside of Taiwan. More important, however, nearly two-thirds of all Japanese took up residence in cities and presumably stayed there (see Table 25). They found an urban setting most advanta-

MIGRATION AND THE GROWTH OF CITIES TABLE 25 PROPORTIONS OF TAIWAN'S POPULATION OF EACH ETHNIC GROUP

Living Inside the Nine Major Cities in Taiwan, and in Taipei City 1920

1940 Sex Ratios

Per Cent of Total Population

Sex Ratios

10.5 60.2 68.4

106 119 345

14.7 65.6 59.8

104 105 175

3.1 27.5 38.6

102 117 364

4.0 29.3 26.8

99 100 166

Per Cent of Total Population Nine Cities Combined

Taiwanese Japanese Other Taipei City Alone

Taiwanese Japanese Other

Source: Census of 1930, Statistical Tables, Table 1.

geous for conducting the new activities that they brought with them to Taiwan. Japanese proceeded to develop Taiwan's cities to pro­ vide an appropriate environment of modern banking facilities, cen­ ters of government administration, reservoirs of professional skills, and headquarters of private business. Personal comforts, too, were not neglected. Running water, sys­ tems of sewage disposal, gas, electricity, and cinemas first appeared in these cities, and did not spread far beyond them. For those who could afford the time and expense, life in Taipei was graced by a nearby golf course and several fashionable vacation resorts. Jap­ anese were energetic in constructing large public buildings conspic­ uously of Western architectural styles, thoroughfares and parks for their gratification and diversion. The section of Taichung City that was occupied by Japanese was referred to fondly as "little Kyoto."13 Cities understandably were attractive places for Japanese to reside. The Japanese affinity for urban life was exceeded only among the mainland Chinese that were found in Taiwan.14 The explanation in their case is somewhat different. These people were admitted to the island as small merchants or as laborers in many of the modern enter13 OPNAV, Civil Affairs Handbook (1944), Taichu Province, p. 149 ff. These made up almost all of the category "Other Nationalities" in Table 25 and in published census tables.

MIGRATION AND THE GROWTH OF CITIES

prises that flourished wherever there were Japanese. In this respect they were distinct from Taiwanese, as their occupational composition indicates (see Table 16). Otherwise they merged in language, cus­ toms, and sympathies most closely with the urban population of Taiwanese. The Chinese tended to remain in cities or commercial towns. The authorities held them under suspicion and close scrutiny, and often confined them by labor contract to menial positions. On the whole they did not become permanent residents of Taiwan; by their sex ratios they appear as a relatively transient group (though some of them intermarried with Taiwanese). Many were merely seasonal workers who remained in Taiwan for only part of the year. The ethnic composition of the urban population was, like its in­ crease, a product of migration from outside cities rather than growth from within. Different cities did not all exert the same attraction for people according to ethnic background. Variations between cities in ethnic composition (see Table 26) were scarcely a matter of TaiTABLE 26 ETHNIC COMPOSITION OF MAJOR CITIES IN TAIWAN

Per Cent of Total Population in Each Ethnic Group 1920, 1930 and 1940 Taiwanese Cities"

Taipei Chilung Hsinchu Taichung Changhua Tainan Chiayi Kaohsiung P'ingtung Total

Others

Japanese

1920

1930

1940

1920

1930

1940

1920

1930

1940

66.4 70.3 90.7 70.5 97.0 81.1 83.8 74.1 84.2

64.1 68.4 87.4 72.8 93.6 80.5 82.3 72.2 82.8

67.9 69.5 86.3 78.9 94.0 86.0 87.4 79.8 85.8

27.8 24.8 8.3 28.0 2.3 15.9 14.5 24.1 14.2

29.4 24.7 11.3 24.4 4.5 15.8 14.8 24.6 13.8

28.1 24.8 12.7 19.8 5.1 11.7 10.8 18.6 12.3

5.8 4.9 1.1 1.5 .7 3.0 1.7 1.8 1.6

6.6 6.9 1.2 2.8 1.9 3.7 2.9 3.2 3.4

4.1 5.7 1.0 1.3 .9 2.3 1.8 1.6 1.9

20.6 21.5 19.2

3.5

4.5

2.8

75.9 73.9 78.0

0 Boundaries of all cities have been adjusted to those of 1935. Sources: 1920 Population Register, Table 2. Census of 1930, Statistical Tables, Tables 1 and 2. Census of 1935, Statistical Tables, Table 1.

wanese being drawn less into some cities than into others, for they comprised the broad population base in every one. These differ-

MIGRATION AND THE GROWTH OF CITIES

ences are to be explained primarily by the preferences of Japanese for certain cities as opposed to others.15 We may thus regard the proportion of Japanese in the population as the major element in the variation between cities. The ethnic composition of each city was remarkably stable from 1920 to 1940, though almost all doubled in population. Taipei, Chilung, Taichung, and Kaohsiung were all more than one-fifth Japanese at each census date. This is least surprising in the case of Taipei, for it was the ad­ ministrative hub of the entire island, and a thriving center of foreign trade as well. Chilung because of its natural harbor became estab­ lished early as the principal port of access to and from Japan. It kept this position even after the harbor no longer handled heavy traffic, by possessing some of the key industries intended to complement Japan's war effort. Taichung was in the fertile middle section of the island, and prospered enormously from the collecting, processing, and marketing of agricultural produce. Kaohsiung was the center from which the Japanese mounted their most extensive efforts at modern­ ization: in the city itself and its immediate environs were concen­ trated the major processing industries of the island. Large sums were spent in improving its harbor to facilitate the passage of this new production out of the island.16 The remaining five Municipalities were more typically Chinese in atmosphere, with the alien features of Japanese presence kept more in the background. Whatever their types, these cities of Taiwan were the most important places where Japanese residents mingled with Taiwanese. Japanese nationals were scattered over the rest of the island in administrative and technical posts, public and private, but only in these cities did they associate regularly with Taiwanese. Only there did many Taiwanese who were non-farmers come into frequent contact with Japanese. Consequently, one would expect that the urban environment was the principal channel of modern influence. Even in these cities there were certain obstacles in the way of such contact. The barrier was not entirely one of language, for Taiwanese soon acquired the rudiments of the Japanese tongue through school15 Mainland Chinese varied almost as much as their share in the total population by city. But they were so few, and their stay in Taiwan so impermanent, that they cannot be considered a very important part of even the urban population. « Kaohsiung City was also built up as a base for southward expansion in the Pacific Ocean, but did not have the opportunity to fulfill the promise of this investment.

MIGRATION AND THE GROWTH OF CITIES

ing and practice. Japanese did many things to discourage free inter­ course between the two groups, and raised exaggerated fears of ani­ mosity from the Taiwanese among their own kind.17 The Japanese maintained high levels of living in Taiwan. Their wealth did not create the separation between the two groups, but helped to enlarge it. The opulence of Japanese was more than simply a matter of superior income (which the Taiwanese might have emu­ lated); it was also founded on privilege. For example, wherever possible they lived apart in special por­ tions of each city. These sections were arranged so as to provide the facilities that they were in a position to command—Japanese shops, entertainment, and cleanliness, one of the greatest of all luxuries in Asia. Influential Japanese strove to put themselves in such surround­ ings, whereupon they allowed poorer sections to develop as Tai­ wanese residents could best afford. Owing to this combination of jealous care and neglect, the central districts of most cities were elaborately laid out with efficient street patterns and handsome land­ scapes—but with residential finery and modern public buildings concentrated in a few blocks and the indescribable slums that Asia often produces spread over the remainder.18 It is easy to see the nature of residential segregation in Taipei City, where it was carried farthest (see Fig. 20, based on 1935 census data). There were two large concentrations of Taiwanese, at the north and the southeast corners. These were the sites of the original pair of settlements that grew into the present city. The space in be­ tween had been gradually built up before 1895 as the seat of Manchu administration for the island. It was razed and replaced by the Jap­ anese (following a destructive typhoon in 1912) to make way for new construction that they planned. The Japanese built this central area themselves, and therefore to their own design. It had a core in which almost none but Japanese eventually lived, surrounded by a 17 One can say that these fears were exaggerated without implying that they were without foundation. By their repressive administration, Japanese invited widespread ill-will in the population. Much of this might have been allayed if they had not on the one hand advertised the benefits of their rule and, on the other, systematically cut off any opportunities for Taiwanese to compete with Japanese on equal footing. This am­ bivalent policy even undermined some of their foremost objectives by impeding their use of the local labor supply in the most effective ways for war production. is The enforcement of health and sanitary regulations held the two groups in com­ plete equality before the law. In looking back at the severity of these measures, Chinese have remembered them as instances of oppression by the Japanese. But by such means the authorities were able to avoid the unhealthful conditions usually associated with urban crowding in Asia. See Hsu Tze-wei (1946), p. 444 ff.

MIGRATION AND THE GROWTH OF CITIES

PER CENT OF JAPANESE IN TOTAL POPULATION, BY WARDS,TAIPEI CITY, 1935

BELOW 50% 0-9.9 10-49.9 ABOVE 50% 50-89.9 90-100

OFFICE OF POPULATION RESEARCH, PRINCETON UNIVERSITY

Fig. 20. Residential Segregation of Japanese in Taipei City.

substantial area where Japanese residents predominated. Around the fringes of this modern district, other Japanese (reportedly those of lower income and social status19) lived more thinly mixed with Tai­ wanese. Almost one-half of the city's surface was taken up by poorer districts composed entirely of Taiwanese dwellings, shops, and open markets. There is also another way to see how the separation of living quarters kept the two groups apart—in terms of persons rather than the space involved. Table 27 illustrates this aspect of segregation for 19 OPNAV, Civil Affairs Handbook (1944), Taihoku Province, p. 133. A more com­ prehensive account of the growth of Taipei City, in relation to the surrounding region, is given in Ch'en Cheng-hsiang (195a).

MIGRATION AND THE GROWTH OF CITIES TABLE 27 RESIDENTIAL SEGREGATION IN TAIPEI CITY, 1925 AND 1935

Taiwanese and Japanese Population Distributed by the Proportion Japanese in Each "Ward" 1935

1925 in Each Ward

Number

Per Cent of Total

Number

Per Cent of Total

Taiwanese

0- 9.9 10.0- 49.9 50.0- 89.9 90.0-100.0 Total Taiwanese

94,551 25,681 7,964 606 128,802

73.4 19.9 6.2 0.5 100.0

3,069 11,462 24,926 13,884 53,341

5.8 21.5 46.7 26.0 100.0

109,309 48,815 16,508 384 175,016

62.5 27.9 9.4 0.2 100.0

Japanese

0- 9.9 10.0- 49.9 50.0- 89.9 90.0-100.0 Total Japanese

3,640 18,173 48,775 11,116 81,704

4.5 22.2 59.7 13.6 100.0

Sources: Census of 1925, Statistical Tables, Table 2. Census of 1935, Statistical Tables, Table 2.

Taipei City in 1925 and 1935. Here both Taiwanese and Japanese have been grouped according to the degree of concentration of Jap­ anese in each "ward." While the non-Japanese wards held the bulk of Taiwanese, there was some considerable area of overlap where the two groups were relatively mixed. From the evidence of this table, many Japanese who were arriving in Taiwan were forced to go to residential areas that contained some Taiwanese residents. This is shown by the relative gain of Japanese in the sections that were not purely Japanese, but where Japanese residents were in the majority. With its members increasing, the Japanese community eventually had to mix more widely with Taiwanese, since the areas that were already more than 90 per cent Japanese could not absorb any new­ comers. They moved into wards with only the "better element" of Taiwanese, however, and managed to keep this mixture to a mini­ mum. The proportion of Japanese living in predominantly Tai­ wanese wards actually fell off very slightly. From the Taiwanese side, moving into mixed sections was carried somewhat further. Taiwanese were beginning to spread into wards that were not purely Taiwanese, sections closer to the center of the

MIGRATION AND THE GROWTH OF CITIES

city which they had occupied together with Japanese. Because there were more of them, and because they had been more concentrated among their own kind before, the Taiwanese moving into these areas contributed more to the mixture than the Japanese. The data here do not signify that it was very extensive. But it makes some difference which of these two aspects of segregation we choose to use as a standard of comparison. This trend in the segregation of urban residents was of some sig­ nificance in the other cities during the years when they were growing rapidly, for the system was put under strain by the continuous ar­ rival of new immigrants. On the whole, the Japanese were successful in housing most of their people in residential wards that were pre­ dominantly Japanese, and well-to-do Taiwanese were moderately successful in entering sections containing some Japanese, but not into the sections most prized by the Japanese. For Japanese, no other city offered living quarters and facilities so exclusive as those of Taipei (see Table 28). In every one of these six other Municipalities they lived more widely intermixed with Tai­ wanese (though not necessarily at the same levels of comfort). KaoTABLE 28 RESIDENTIAL SEGREGATION IN SIX CITIES OF TAIWAN, 1935

Taiwanese and Japanese Population Distributed by the Proportion Japanese of Each Ward Fer Cent Japanese in Each Ward

Chilung

Hsinchu

56.6 34.2 8.9 0.3

68.5 30.7 0.7 0.1

52.7 42.7 4.5 0.1

100.0

100.0

100.0

5.7 35.6 50.9 7.8

16.4 54.3 16.6 12.7

6.8 53.0 34.9 5.3

100.0

100.0

100.0

Taichung

Chiayi

Kaohsiung

50.0 48.3 1.7 0.0

58.2 41.5 0.3 0.0

56.1 36.2 7.7 0.0

100.0

100.0

100.0

11.0 69.9 11.1 8.0

15.7 80.0 4.3 0.0

7.9 38.6 53.5 0.0

100.0

100.0

100.0

Tainan

Taiwanese

O- 9.9 10.0- 49.9 50.0- 89.9 90.0-100.0 Total Taiwanese

Japanese

0- 9.9 10.0- 49.9 50.0- 89.9 90.0-100.0 Total Japanese

Source: Census of 1933, Statistical Tables, Table 2.

124

MIGRATION AND THE GROWTH OF CITIES

hsiung and Chiayi had not a single ward with as high as 90 per cent of its population Japanese. For Taiwanese, on the other hand, Taipei consisted neither more nor less of mixed wards than the other cities. In all, at least nine-tenths of the Taiwanese population of every city stayed in districts in which less than a majority of the inhabitants were Japanese. The Taiwanese in Cities

Of all Taiwanese, so few went to cities that one is prompted to ask whether there were not some special factors present restricting the size of this migration. The effects of such a selective process over many years should produce differences in the characteristics of per­ sons living in cities and of those living outside. Since few if any newcomers had an opportunity to take up agri­ culture in the cities, a prominent feature of cityward migration was a movement to new work. To the extent that it was, one would ex­ pect that the migrants were the types of persons who usually under­ took new occupations in Taiwan—for example, men rather than women, primarily in the ages of young adulthood. If the migrant population had been limited to these categories, cities with large proportions of migrants in their populations must have had dispro­ portionate numbers of young men as compared with the rest of the island. Such a tendency has characterized rapid urbanization else­ where. In addition, cities frequently offer opportunities for employment that are considered superior to those found in the countryside. This is especially true of many agrarian societies, where rural poverty sometimes reaches extremes that make the meanest of urban jobs seem attractive. Where such conditions exist, movement to cities very often serves to drain off large numbers of men in young working ages who have lost or overcome their attachment to farming. In 1935, these effects of age and sex selection were hardly per­ ceptible in Taiwan's urban population taken as a whole. The age composition of all their males combined was nearly the same as the rural population (Table 29). Surprisingly, urban females likewise had almost the same age structure as those outside of cities. Hence, there could have been no sweeping dislocation of persons of different age and sex groups as a result of urbanization in Taiwan. There are special circumstances in which it is not hard to imagine some movement of single individuals in an agrarian society; this has

MIGRATION AND THE GROWTH OF CITIES TABLE 29 AGE AND SEX COMPOSITION OF TAIWANESE, INSIDE AND OUTSIDE OF NINE MAJOR CITIES, 1925 AND 1935

Per Cent of Total Population in Selected Age Intervals, Male and Female Males Outside

Age

Inside

0-14 15-39 40-64 65+ Total

36.5 44.1 17.8 1.6 100.0

40.2 40.6 17.4 1.8 100.0

0-14 15-39 40-64 65+ Total

41.0 41.6 15.9 1.5 100.0

43.2 38.5 16.5 1.8 100.0

Females Males per 100 Females Outside Inside Inside Outside 1925

37.5 39.5 19.5 3.5 100.0

40.0 37.6 18.8 3.6 100.0

101 116 95 47 104

105 113 96 52 104

42.9 37.0 16.8 3.3 100.0

103 110 97 50 103

103 107 101 57 103

1935

41.1 39.0 16.8 3.1 100.0

Sources: Census of 1925, Statistical Tables, Table 3. Census of 1935, Statistical Tables, Table 7.

been known to occur in the form of contractual servitude by chil­ dren, of young men (though not frequently young women) shifting to non-agricultural employment, and of movement by old persons to reside with emigrant kin. But it is inconceivable that these deviant events might have occurred jointly in chance coordination so as to reproduce the rural age structure so closely in cities. Such a migra­ tion as this could have been possible only if many people travelled to cities in family units and not as isolated individuals. A great many migrants therefore must have moved merely because they were mem­ bers of migrant kinship groups.20 This situation did not hold with equal force in all cities. Varia­ tions in composition among separate cities were as great as the dis­ tinction between city and non-city. Assuming the age group 15 to 39 among males to be the most responsive to new opportunities for 20 Estimates for the age composition of the migrants lend weight to this supposition. Survivors to 1935 of people who were already in cities in 1925 can be estimated at every age by subtracting the deaths that one would expect of them; the amount by which these survivors fail to account for the people actually found in cities in 1935 gives a fair idea of the persons gained through migration at each age. Though migrants meas­ ured in this way were somewhat concentrated in young adult ages, there were also im­ pressive numbers in younger and older ages.

MIGRATION AND THE GROWTH OF CITIES TABLE 30 DIFFERENCES IN AGE AND SEX COMPOSITION OF TAIWANESE BY CITY, 1935 Growth of City: Ratio of Taiwanese Population

Per Cent 15-39 of all Taiwanese Males

Sex Ratios

15-39

Sex Ratios all Ages

Taipei Chilung Hsinchu Taichung Changhua Tainan Chiayi Kaohsiung P'ingtung

41.8 42.4 38.2 43.2 39.0 40.6 42.2 42.8 43.5

106 112 101 115 102 110 112 120 118

98 103 100 107 100 105 105 110 108

2.05 2.07 1.76 2.92 1.67 1.97 2.58 3.56 2.55

Total, All Cities

41.6

110

103

2.22

Cities

1940/1920

Source: Census of 1935, Statistical Tables, Table 7.

employment, we can use this group to summarize the differences (see Table 30). The cities of rapid growth—i.e., with relatively large proportions of immigrants—were the cities with more young men. Cities that grew faster did so in part by attracting more men in young working ages, and thus separated many of them from normal family life. Significantly, the age composition of females was nearly the same in all the individual cities, and thus was similar in each to rural communities. The non-rural character of Municipalities in Taiwan, therefore, did not of itself make for non-rural age structure, but the same factors that were behind the more rapid growth of some cities tended to draw selectively from the population of different ages outside cities. One should always bear in mind that the greatest of the dif­ ferences in Table 30 are moderate when considered alongside so­ cieties in which urban growth has been associated with sweeping modernization. Something in addition can be learned about the particular charac­ teristics of separate cities from the nature of economic activities in each. One single census table of persons with occupations was pub­ lished in 1930 for individual cities, with the males of all ethnic groups lumped together. Cities were farthest removed of any civil

MIGRATION AND THE GROWTH OF CITIES TABLE 31 INTER-CITY DIFFERENCES IN OCCUPATIONAL COMPOSITION

Per Cent of Occupied Males in Selected Types of Occupation in Each of Seven Municipalities, 1930 All Nationality Groups Cities

Taipei ChiIung Hsinchu Taichung Tainan Chiayi Kaohsiung Total Seven Cities

Manu­ facturing

Commerce

Trans­ Govern­ portation ment

Agri­ culture

Other Occupations

26 16 21 23 28 22 24

30 21 25 28 28 29 19

10 23 10 10 13 10 20

18 9 13 19 18 12 11

6 4 20 14 5 20 13

10 27 11 6 8 7 13

24

Tl

13

15

9

12

Source: Census of 1930, Statistical Tables, Table 40.

divisions from rural conditions. Nevertheless, large blocs of Taiwan's urban population still supported themselves by farming. Only three Municipalities in 1930 had as few as one-twentieth of their working males engaged in agriculture. It is small wonder that these cities had failed to show symptoms of advanced urbanism in other respects. Other types of employment in these cities also had certain special features. Government administration in Taipei, transshipping of seaport cargoes in Chilung and Kaohsiung, and sugar processing in Taichung and Tainan were reflected as important activities in their share of all male employment. These figures, pertaining to 1930, reveal nothing of the intensified industrial activity in these cities in the succeeding decade. They are the only available indication in demographic terms of the peculiarities of individual cities. In any event, we are less interested in these peculiar elements than in the relations between types of employment and urban settlement among Taiwanese. Since Table 31 does not show Taiwanese separately, it is affected by differences in the proportions of Japanese and in their work among the various cities. Hence the information, while illu­ minating, is less useful than one might wish. More information is available for the population of all seven cities combined. By means of some estimation, we can find the occupational

MIGRATION AND THE GROWTH OF CITIES TABLE 32 OCCUPATIONAL COMPOSITION IN CITIES OF TAIWAN BY ETHNIC GROUP, 1930"

Distribution of Occupied Males by Occupation and Nationality Taiwanese

Agriculture Fishing Mining Manufacturing Commerce Transportation Govt. & Prof. Other

A. Per Cents by Occupation 0.9 13.1 2.2 2.1 0.2 2.1 22.8 18.6 29.1 19.6 10.2 13.7 44.4 6.4 3.9 10.7

Total Occupied

Other Nationalities

0.2 0.5 1.1 43.8 27.8 19.1 3.3 4.2

100.0

100.0

100.0

54.4

58.0

76.8

B. Per Cents by Ethnic Group 2.4 97.5 27.2 70.8 89.9 3.8 19.0 65.0 73.2 17.8 69.1 18.5 28.2 70.0 84.7 11.1

0.1 2.0 6.3 16.0 9.0 12.4 1.8 4.2

24.0

8.6

Total Occupied Per Cent Occupied of All Males Agriculture Fishing Mining Manufacturing Commerce Transportation Govt. & Prof. Other

Japanese

67.4

e Includes the seven cities of Municipal rank in 1930. Source: Estimated from Census of 1930, Statistical Tables, Table 39.

composition of each ethnic group separately (Table 32).21 Taking account only of males with occupation, these data represent the urban Taiwanese at less of a disadvantage next to the Japanese than was true over the entire island. Of course, one must remember that Tai­ wanese (and Mainland Chinese) in the same type of work with Jap21 The necessity to "estimate" arises because the occupational data by ethnic groups were given in "Occupation" units (i.e., persons with a principal occupation plus per­ sons with a secondary occupation), and counted some persons twice. In the tables shown here, the ethnic composition of "Occupations" has been applied to the total of persons (males) in each occupation.

MIGRATION AND THE GROWTH OF CITIES

anese usually had subordinate positions. Nevertheless, it seems con­ clusive that by moving into cities Taiwanese were able to narrow the gap that separated their forms of livelihood from those of the Japanese. The ethnic composition of occupied persons (appended to Table 32) indicates how the differences in occupational structure among these ethnic groups were weighted in the urban population. Tai­ wanese, of course, predominated in all lines except in government service and professional activities. Keeping these coveted positions in their own hands was one of the ways in which Japanese were able to live side by side with the rest of the population and avoid the pressure of competition. Chinese nationals ("Other Nationalities" in Table 32) in these cities also appeared in a few types of work far more than in the work­ ing population as a whole. Their specialties were of an opposite kind: many worked as laborers in small manufacturing enterprises and as longshoremen, accounting for their large share of all "Manufactur­ ing" and "Transportation" workers. TABLE 33 OCCUPATIONAL DISTRIBUTION OF TAIWANESE MALES REPORTING AN OCCUPATION, INSIDE AND OUTSIDE MAJOR CITIES, 1905, 1920, 1930 1905 Insideti Outside

Agriculture 2.3 Fishing .6 Mining Ί L 28 4 Manufacturing Commerce 37.7 Transportation 10.4 Govt. & Prof. 5.6 Other 15.0 Total Occup.

100.0

1920 Inside\ Outside

1930 Insidej Outside

6.4 1.9 1.7 9.6

13.2 1.2 f 1.1 \ 27.1 28.9 11.2 6.7 10.6

71.9 2.3 1.4 6.9 6.4 2.8 1.9 6.4

13.1 2.1 2.1 22.8 29.1 13.7 6.4 10.7

74.0 2.0 1.3 5.9 8.2 2.8 2.3 3.5

100.0

100.0

100.0

100.0

100.0

72.1 2.9 5.4

4 Comprises the cities of Taipei, Taichung, and Tainan. The boundaries for Taipei in this tabulation were smaller than the city as subsequently organized. f Cities of Taipei, Taichung, and Tainan, with the same boundaries used at later census dates. I Cities of Taipei, Chilung, Hsinchu, Taichung, Tainan, Chiayi, Kaohsiung. Sources: Census of 1905, Detailed Tables, Prefectural Volume, Table 17. Census of 1920, Prefectural Volume, Table 16. Census of 1930, Statistical Tables, Table 39.

MIGRATION AND THE GROWTH OF CITIES

Finally, the role of migration in social change can be clear only if we can make comparisons at different points of time. Occupational profiles of city and non-city working males are available in not quite the same form for three separate dates. The census reports of 1905 and 1920 each included a table of occupied persons in three cities (Taipei, Taichung and Tainan). Placed next to the corresponding figures in 1930 (Table 33), these early data are reasonably similar in all three years. Not all of the same cities are included, and so the occupational distributions for cities alone do not stand close comparison with each other. One's attention is therefore drawn to the structure of occupa­ tions outside the cities involved. In Table 33, it appears that by re­ moving the small group of urban working persons from the total population of Taiwanese, one in effect also removes most of the di­ versification in employment over time, and consequently a large share of the social changes that occurred in Taiwan. In other words, the new activities that resulted from Japanese occupation remained sealed up in the few urban centers. The Taiwanese population out­ side of these cities remained essentially as agrarian as at the begin­ ning of Japanese rule.22 Even if as centers of innovation we did not consider Taiwan's cities important, a study of migration would have made them so. Per­ manent migration was rare among Taiwanese; but of those who migrated, a large proportion went to a few cities. Even though these cities were not growing fast enough to enlarge their share of all Tai­ wanese very substantially, they did the greatest part in redistributing the population. The real significance of these shifts of residence arose from the fact that they were also shifts in activities. For the most part with­ drawal from agriculture meant exit from rural communities. People departed every year to large towns or cities that had some provision for more diverse kinds of livelihood. It was primarily in these cen­ ters that the Japanese fostered modern enterprises and developed a need for labor. Migration was one of the forms of the response to 22 Obviously, since more cities are represented in the 1930 figures, the figures them­ selves are biased and there is some chance o£ bias in the conclusion that we have drawn. By taking occupied men of all ethnic groups, the same comparison can be made based on the same three cities at all dates. This is possible because the table for all ethnic groups was published for each city separately (cf. Table 31, above). The result is the same, in that the data indicate no change in occupational structure outside even these three cities.

MIGRATION AND THE GROWTH OF CITIES

these new opportunities for employment. It steadily brought more Taiwanese into contact with the modern side of the Japanese eco­ nomic program. Yet even though movement to the cities did widen the horizons of a few Taiwanese, it was surrounded by circumstances that tended to minimize this effect. The familial character of the migration made the break between country and city less sharp than it might have been. Removal to a city did not expose migrants very readily to what­ ever new influences were there. These were all dominated by Jap­ anese, who kept most of their innovations to themselves and erected barriers against free intercourse with the rest of the population. Urban growth therefore did not promote general social change among the Taiwanese. It removed into cities those who were to feel the strongest effects of contact with these alien elements. Migration thus helped to absorb the modern aspects of economic development so that they did not intrude into the affairs of rural communities.

CHAPTER VI

Public Health and the Risks of Death I N the preceding pages, we have given attention to various features of intercourse between Taiwanese and Japanese in Taiwan. Most of the evidence has been in agreement on one point—the policies of Japanese as rulers of the island failed to make any substantial dif­ ference in the social structure of its people. The more it seemed to change in some particular, the more it turned out to be essentially the same thing. As a matter of official design many innovations were concentrated in the hands of a favored few. Even when it came to implementing their aims, Japanese found ways to meet the popula­ tion on its own terms. Much of their success in colonial administra­ tion was due to the ability to make the social system serve their purposes without changing its form. The most significant exception was their effort on behalf of the health of the inhabitants. Japanese have never received full credit for their achievement in this realm, either on the part of the Tai­ wanese (who resented the intrusions that it involved in their lives) or from Western students of population and public health (who mostly ignored or belittled it). The "benefits" of rationalizing Tai­ wanese agriculture may be a matter of dispute, for it enormously en­ riched a few Japanese at the same time that it broadened standards of consumption in the population. The motives of Japanese in intro­ ducing to Taiwan some of the advantages of scientific discovery from the outside world may likewise be suspect, for they managed to keep control over most of the fruits of improvement. But there is no doubt that in the field of public health Japanese rule added many years to the expected lifetime of Taiwanese. This, in an environment of opinion that looks with approval upon any postponement of death, deserves to be given adequate attention.

Early Health Conditions in Taiwan

Prior to 1895, Taiwan was renowned as an unhealthful place to live. It presented special hazards to survival in addition to those which its people had learned to accept in South China. Unusual efforts had been called forth in expanding settlement into the rela­ tively empty area, and these entailed unusual risks of death. Early Chinese migrants to Taiwan had to attend to clearing new lands,

PUBLIC HEALTH AND THE RISKS OF DEATH

and to contend with many types of hardship not found on the main­ land. As they pushed toward the island's foothills, they also suffered the consequences of many bloody incidents of violence that they provoked among aboriginal tribesmen. Even in the lowlands, where activities were comparatively settled and routine, life itself was a very uncertain thing. From their en­ vironment people were subjected to attack by several kinds of fatal sickness. Through certain traditional practices they unwittingly ex­ posed themselves to many others. The semi-tropical climate of Tai­ wan favored the spread of a multitude of disease-producing agents. Accustomed methods of cultivating the land—especially those hav­ ing to do with the conservation of human wastes as fertilizer—en­ couraged the transmission of these organisms to humans.1 Through conditions such as these, the population harbored many body parasites and other sources of infection. As a result, many people passed their lives chronically enfeebled, and others remained con­ stantly exposed to the onset of illness. Certain types of disease often spread over the island, restrained only by the isolation of many localities and the difficulty of travel. From time to time, some would grow to spectacular epidemics in which thousands perished. Disease even proved to be one of the best defenses that Taiwan had against attack by the troops of Western powers, and delayed the Japanese in taking control of the island. Plague was the foremost among these epidemic diseases.2 Taiwan was repeatedly swept by attacks of this illness, with heavy fatalities. An uncontrolled traffic of shipping from the mainland and from else­ where in Asia, much of it illicit, made it easy to introduce infected rats and persons. Conditions of knowledge and sanitation in Taiwan were wholly inadequate to cope with the problem.8 Cholera was the second most important disease that occurred in epidemic proportions. Lacking knowledge or facilities to control the spread of cholera, Taiwanese sometimes suffered the full course of epidemics throughout a whole season without being able to take counter-measures.4 Cholera also was well known and feared in the 1 For an able popular account of some of the connections between custom and the "food cycle" that propagates these organisms, see Winfield, G. F. (1948), p. 112 ff. 2 See Hsii Tze-wei (1946), p. 444. 3 For some of the futile steps adopted against plague by the people, see Campbell, W. (1915). P-77· Drinking water was usually taken from the rivers or from shallow wells, making them easily subject to pollution. See Takekoshi, Y. (1907), p. 285.

PUBLIC HEALTH AND THE RISKS OF DEATH

early Western contacts with Taiwan. It all but wiped out some of the military expeditions sent there as tentative colonial ventures by European powers, and later helped to persuade the Japanese to em­ bark on large-scale measures of control. Malaria was ordinarily endemic in Taiwan. The moist heat, and the practice of admitting water into fields for irrigation in pools that were left to stagnate, helped to nurture the mosquitoes that spread the disease. Occasionally malaria broke out in epidemic form. It was likewise impossible to control, but was simply allowed to run its course. Even in the more fortunate years it served to undermine the strength of Taiwanese during much of the year, and probably also their resistance to other sickness. In addition to these major diseases, Taiwanese were widely afflicted with the more common debilitating ailments that tend to shorten life and lead to complicating illnesses in Asia. Enteric diseases and internal parasites were always prevalent, as in China. Tuberculosis and venereal disease were later found to be widespread, and pre­ sumably had existed also at this earlier period. Superficial skin and eye maladies were frequent. Taiwan's position as a transshipping point in the opium trade introduced a plentiful amount of the drug for local consumption, which impaired the health of the inhabitants still further. In the census of 1905, it was found that nearly one-fifth of all Taiwanese men over age 30 were known users of opium, and there were probably others who managed to conceal the fact. As external trade increased, surreptitiously and openly, Taiwan became exposed to a wider variety of diseases and a greater number of trans­ mitting agents from outside. At the same time, the official policy of seclusion also kept out the techniques that the West was evolving to deal with them. These baneful health conditions added new hardships for the people who migrated to Taiwan. The risks of death thus engendered became a formidable feature of everyday living. Moreover, under the pressure of poverty many families apparently eliminated certain unwanted offspring, either through exposure or deliberate neglect. As a result, a large proportion of those who were born into the popu­ lation never survived to adulthood. The dangers to life were such that probably Taiwan gained few if any people through natural in­ crease; very high fertility was needed merely to offset the losses through death.

PUBLIC HEALTH AND THE RISKS OF DEATH

The Japanese and Public Health The Japanese military forces that first set out to subdue Taiwan met no firmly organized resistance among the people, but were nearly overcome by sickness in their own ranks before taking posses­ sion of the island. Out of an undisclosed total number of troops taking part in the campaign, the Japanese acknowledged only 154 combat fatalities, but over 4,000 died of disease within the first sev­ eral months, and 27,000 had to be hospitalized (most of them in Japan).5 Epidemic illness did not, like poverty, stay confined to the Taiwanese population and leave the new rulers unmolested. Mos­ quitoes that carried malaria, for example, did not observe this dis­ tinction. The Japanese liking for raw fish was said to make them even more vulnerable than Taiwanese to the parasites that infested the food. Among the garrison alone, annual deaths of more than 40 per thousand were reported during the first decade of occupation, probably a higher figure than in the adult population of Taiwanese. In order to make this region of the world safe for imperialism, the Japanese found that they had to take drastic measures to curb the spread of disease. They started almost at once to devise ways of counteracting the most serious sources. The steps that they adopted are instructive, for the result was a marked success with only a modest outlay. The cam­ paign against sickness was launched when scientific knowledge as to the nature of these diseases was by present standards still rudimentary. Its weapons were crude and sometimes misguided. Moreover, there were many other goals of administration that competed with public health for attention. The authorities were never in a position to di­ vert large blocs of funds to make their program elaborate or mu­ nificent. It lacked the grandeur of projects sometimes contemplated by those who draw up plans for "backward" areas today. The whole approach of the Japanese to the problem of disease in Taiwan was an austere one. They avoided every possible expense in equipment and physical facilities, and depended on intensive use of all the ad­ ministrative resources at their command. In slightly over a decade the authorities almost eradicated plague as a major disease, through strict vigilance for suspected cases of in­ fection, rigorous quarantine regulations, and careful inspection of incoming ocean cargoes for infected rats.6 Cholera was similarly pre5 Davidson, James W. (1903), p. 354. Naito, Hideo (1937), p. 84 ff.

6

PUBLIC HEALTH AND THE RISKS OF DEATH

vented from spreading in the population by incessant search for afflicted persons, and quarantine of those cases that were found. In­ oculations gradually were developed into an indispensable preventive measure against both of these diseases. The cities, where crowded and unsanitary living conditions in­ vited some deterioration of health standards, were made less favor­ able to cholera by piping in water from mountain streams and from specially constructed wells. In nearby localities, regulations were issued—and enforced—requiring villagers to tap sub-surface sources of water with deep wells that were kept covered.7 Contamination of food supplies in the larger cities was minimized by crude provision for sewage disposal—a few septic tanks and public sewers were built, but most municipal authorities merely supervised the removal of sewage by cart. Agencies of the Sanitation Bureau regulated food handling and butchery in the public market places. Malaria had to be attacked indirectly, and hence proved more stubborn to keep in check. It remained a source of widespread ener­ vation and an important cause of death even during the 1930's, when measures for its control had become most effective. There were then no very satisfactory ways to kill mosquitoes or to check their breed­ ing under conditions in Taiwan. The government drained many swamps, but the irrigation procedures necessary for the island's food supply depended on open water storage and flooding of fields.8 The localities most frequently affected soon became known, and the au­ thorities relied on detection and compulsory treatment of cases that were discovered. They sent special medical personnel to these dis­ tricts, backed up by police power, and forced all suspected individuals to undergo blood tests and compulsory treatment with quinine. In the later 1930's, over three million examinations were regularly be­ ing performed each year. By unremitting and energetic insistence on these devices, the Japanese managed at least to prevent any large epidemic outbreak of malaria during their stay in Taiwan. With these few outstanding dangers under control, the principal remaining causes of ill health and death were respiratory ailments and enteric disorders. As the former ceased to produce epidemics, these more common types of sickness came to account for an increas7 U.S.

War Department. Technical Bulletin (TB Med 30) (1944), p. 2. It is moreover alleged that certain types of mosquitoes bred in streams of the in­ terior foothills. See U.S. War Department, Technical Bulletin (TB Med 30), p. 9. 8

PUBLIC HEALTH AND THE RISKS OF DEATH

ing share of all deaths.9 They were diseases that could readily be avoided by individuals whose livelihood did not continually expose them to contact with likely sources of infection, and whose income permitted them to observe simple precautions. Most Japanese were consequently protected by the nature of their occupations from the common forms of fatal illness in Taiwan. These diseases did not jeopardize their possession of the island as had epidemics in the past. Even so, the Japanese conducted among the Taiwanese population a wide program of education and instruction in some of the simpler techniques of contending with these sources of ill health. They issued countless sanitation rules, and utilized much of their apparatus of government to enforce their observance.10 Administratively, the health and sanitation functions of government were coordinated in several departments. In practice most of the regulations were put into effect under the authority of the police. Some trained special­ ists participated, bringing elements of Western medicine to Taiwan where formerly there had been none except for scattered and un­ coordinated efforts of missionaries. The level of skill among the personnel at their disposal suffers by comparison with that of modern Western countries, and it probably also fell below the standards of the more restricted public health work in some of the tropical dependencies of European powers. The strength of the system established by the Japanese, however, was in its organization and generality, and not in its equipage. It was a com­ prehensive program rather than a piecemeal offering of the most elaborate medical treatment.11 Though the physical facilities of sani9 For reasons that we shall come to presently, the statistics on causes of death are hard to interpret in Taiwan. They show clearly, however, a decline in malaria and the epidemic diseases, and a relative increase in deaths due to these other more insidious ailments. 10 Indeed, it sometimes almost appeared that they used these objectives to cement their control over the population. The enforcement of public health regulations pro­ vided an excellent occasion for the police to enhance their authority over everyday activities. The Taiwanese who protested that these measures were mere excuses to interfere in the private lives of the people probably had a certain amount of justification for their charges. See Hsti Tze-wei (1946), p. 444. 11 The Japanese did not ignore training, but assigned it a special role in their gen­ eral health program. Various medical schools were established for medical instruction of Taiwanese, the first as early as 1902. The staffs of these schools and part of the faculty at Taihoku Imperial University included some competent medical specialists, but these men were retained primarily for research that had no close bearing on public health in Taiwan. Through such limited educational facilities as these and the licensing of local doctors and midwives, Japanese managed to give elementary instruction to many Taiwanese practitioners. Though superficial, the training was distributed widely enough for a striking effect on mortality.

PUBLIC HEALTH AND THE RISKS OF DEATH

tation were confined to some of the major cities, the influence of these broader health measures gradually spread evenly over most of the island. In this way Japanese gained results that were very sub­ stantial through the use of none but crude techniques. The Japanese policy toward opium addiction was less consistent, for here they found welfare goals to be in conflict with profit motives. They made the manufacture of opium a government monopoly, and closely regulated its sale. The announced purpose for this control was to eradicate the use of the drug, but the Taiwanese occasionally complained that officials secretly abetted its consumption. The truth of the matter is not clear, but their success in other tasks suggests that opium addiction could have been stamped out far more quickly if they had made a determined effort to do so. Some of their own records show that new users of opium were being granted licenses as late as the 1920's. It is likely that the Japanese appreciated the effectiveness of addiction in enervating the resistance to their rule among wealthy families, and they made no secret of the steady public revenues that they derived from opium sales.12 The Evidence for Measurement of Mortality

There is not much point in describing the purposes and methods of public health work without having some independent appraisal of results. Such efforts have often been sponsored elsewhere with much fanfare and little accomplishment. The only trustworthy way to view the results of economic and sanitation policies in prolonging life is through some form of measurement; this is likewise the only precise way of comparing the risks of death in Taiwan and in other areas. Fortunately the data that exist to gauge mortality in Taiwan are excellent. The occurrence of many kinds of illness involving danger of contagion was required by law to be reported to the authorities. Partly as a means of avoiding contamination of the living by the dead, and partly just as a means of tightening official control over the reporting of disease, the Japanese also strictly regulated burials and cremations.13 The rigors of enforcing these rules led to conceal­ ment of some illness and evasions by mis-stating exact causes of 12 A careful study of opium addiction in Taiwan and its relation to mortality is con­ tained in Tu Tsung-ming (1935 and 1938). 13 See Hsu Tze-wei (1946), p. 449.

PUBLIC HEALTH AND THE RISKS OF DEATH

death;14 but at the same time they led to registration of deaths that was virtually complete. A person while living had a legitimate civil status only by being entered in an appropriate police register. Deceased, his absence could be legitimized only by removal from these rolls, and his burial could be permitted only by following this prescribed procedure. All of these records were kept by the police or by affiliated agencies. The police were alert and well-deployed to prevent evasions. Since such records were useful to them in surveillance of the population, they employed the pao-chia system to make the registration effective. The people in each locality were thus compelled to accept conformity with these requirements as a matter of their own collective interest. Consequently a person's identity in these records from birth till death was almost never subject to confusion or concealment. This consistency of information in all the population registers has a special significance for demographic research. It means, for example, that each person's death record was matched to the record of his birth, so that the exact number of years elapsed between the two events can be calculated. The same legal identity of a person formed the basis for his entry in every census enumeration. "Age," as a cen­ sus category, was not set down according to an individual's recollec­ tion of the number of birthdays experienced. Instead, the pertinent fact that was entered was his date of birth, and age was ascribed to him in the process of tabulation at the central census office. Age in Taiwanese population statistics was therefore explicitly defined: it was a classification of equal intervals, in which each per­ son ideally had one and only one "true" position at a time. It was possible for age to correspond perfectly in the recording of all three kinds of events of interest here—birth, enumeration at census date, and death. An unambiguous scheme of this sort for reckoning age is especially important in a population of Chinese, because most Chi­ nese compute their ages in a manner that cannot be reconciled to Western usage. Starting with age one at the date of birth, it is cus­ tomary for them to add a year of age with the passage of every (lunar) Once cases of smallpox, scarlet fever, typhoid fever and certain other diseases were discovered, the people were forced to undergo disinfection of lodgings and personal effects that was often destructive. Hence, it has been alleged that diagnoses were de­ liberately falsified to avoid these consequences (see OPNAV, Civil Affairs Handbook, 1944, Shinchiku Province, p. 54; U.S. War Department, TB Med 30, p. 6). A more im­ portant source of inaccuracy was probably (in the case of deaths) ignorance as to the nature of a decedent's illness. Tuberculosis was said to be very prevalent, and yet few deaths were ever diagnosed as due to this cause.

PUBLIC HEALTH AND THE RISKS OF DEATH

New Year thereafter. The Japanese system of ascribing age com­ pletely avoided bias from this source. Moreover, in this system the year of age that a person occupied at any date did not depend on his personal conception of it at the time. Such subjective conceptions of age often do not change through an individual lifetime exactly as "true" age does. Misstatement of age is always a source of error in census procedures that rely on people to report their own ages directly. The recorded ages of people in Tai­ wan at census dates and at death were determined at the date of birth. If the birth had been registered on the proper date, it was next to impossible for the reporting of age in a census enumeration to be incorrect. Even when the date of birth was reported wrongly, there was strong pressure to be consistent later when reporting it in the census and at death. For purposes of constructing life tables, the system thus made an automatic adjustment for these errors by shift­ ing the age at death by exactly the same amount.15 Decedents at a given year of age were registered according to the particular calendar year in which they were born; persons enumerated in a census of the same year at the same age were the survivors of births in that prior calendar year, from January to December. It follows that both registered deaths and census of population belonged to the same group of people. With these two figures, it is a simple problem in arithmetic to compute the number of persons who entered that year of age, the number who died before reaching the next higher age, and the proportion who died out of the first group.16 This propor­ tion is the most important element in the construction of life tables, and it places Taiwan's mortality data in this respect among the most remarkable that exist. The applications of this special feature are discussed in the next section.17 15 The census of 1905 enumerated Taiwanese with birth dates as far back as the year of Meiji Restoration and earlier. Though evidence like family records and land titles were often called upon by the authorities in order to substantiate this information, some inaccuracies must have been present. They could not have been numerous, for they do not appear as "heaping" in census age distributions of people. 16 Of course, the census population must be brought up to December 31 if the census was taken earlier (which has always been the case in Taiwan). Also, the figure for deaths must be re-constituted from the vital statistics reports of two years, since many of the decedents died later than December 31 of the census year. These two details do not alter the principle explained above. 17 In the earlier census enumerations, before 1930, age was ascribed by calendar year of birth. Persons born after January 1 in the census year were placed into the first year of age, those born in the calendar year preceding the census into the second, and so on. To this peculiarity of the system we owe the possibility of making accurate early life

PUBLIC HEALTH AND THE RISKS OF DEATH

A welcome feature of Taiwan's mortality and census data is that they lend themselves to a detailed test for accuracy. First, they can be checked against the number of people counted in one census who died before the next one. For any given cohort of births, if the data are all correct, this should be equal to the number of survivors counted in the earlier minus the number counted in the later census. The censuses of 1905 and 1920 are compared in these terms in Table 34, for males only; the result is almost identical for females. The column of ratios indicates the degree to which registered deaths ac­ count for the difference in the numbers of survivors counted at suc­ cessive census dates. If all records were complete and accurate, the ratios would be unity at every age. If the differences in census survivors are assumed to be correct, then apparently a few deaths went unregistered in certain years. It is obvious that the records were not perfect. As it is, they were so nearly complete that there is little if anything to be gained from trying to make corrections for the discrepancies. Since this is simultaneously a test of census enumeration and vital regis­ tration (and an extremely sensitive test of age reporting), it is not possible to decide whether the discrepancies arose from deaths that were omitted or people overlooked from the earlier census.18 Since mortality in the first year of life is usually measured on the basis of the number of live-born babies, it is necessary also to know how adequate the birth statistics were in relation to death and cen­ sus records. A test for births, similar to that above, is shown in Table 35; the results give an excellent account of the birth registration. The registration of birth entered a child permanently in the police lists, and was an almost inescapable bond during his later life. By this criterion, the chances of evading subsequent controls allied to tables for Taiwanese, for the yearly intervals had the same upper and lower limits as those of the vital statistics. Japanese were apologetic about the deviation of this system from the Western mode of age reporting, and beginning with 1930 "modernized" the system so that its special virtue was eliminated. Starting with the census of that year, census-year rather than calendar-year intervals were used, making the system identical with the Western con­ ventions in design. For dates after 1930 it is no longer possible to relate deaths pre­ cisely to enumerated census population. is There is reason for not putting all the blame on registered deaths, for the area of effective administration was expanded slightly after 1905, and by 1915 and 1920 had brought more people into the registration and census system than the survivors of those enumerated in 1905. This being the case, it is possible that both measuring instruments were nearly perfect, and that the discrepancies arose entirely from the expansion of administration.

PUBLIC HEALTH AND THE RISKS OF DEATH TABLE 34 TEST FOR ADEQUACY OF DEATH REGISTRATION OF TAIWANESE

Males by Year of Birth, Surviving at Census Dates in 1905 and 1920, and Intervening Deaths as Registered Year °f Birth

Survivors 1905

Survivors 1920

1904 1903 1902 1901 1900

48,859 43,432 43,340 41,658 41,910

1899 1898 1897 1896 1895 1894 1893 1892 1891 1890

38,149 40,913 34,183 33,855 35,579

38,257 35,769 36,392 35,523 36,224 32,899 35,409 29,707 29,149 30,148 27,187 27,062 27,945 25,963 24,672 25,479 25,907 27,101 25,254 24,708 27,156 25,257 23,034 25,034 21,484

1889 1888 1887 1886 1885 1884 1883 1882 1881 1880 1879 1878 1877 1876 1875

32,338 32,581 34,122 32,421 31,626 32,164 34,355 35,471 33,697 33,445 36,969 34,750 32,431 35,951 31,335 31,345 29,016 27,669 31,238 27,991

21,230 19,036 18,379 20,263 17,880

Difference

10,602 7,663 6,948 6,135 5,686 5,250 5,504 4,476 4,706 5,431 5,151 5,519 6,177 6,458 6,954 6,685 8,448 8,370 8,443 8,737 9,813 9,493 9,397 10,917 9,851 10,115 9,980 9,290 10,975 10,111

Registered "Completeness Deaths of Death 1905-1920 Registration*

11,602 7,768 6,718 6,087 5,564 5,099 5,257 4,801 4,872 5,177 5,052 5,444 6,224 6,263 6,540 7,273 8,040 8,584 8,587 9,016

.981 .986 1.008 .970 .940 1.088 .952 1.026 1.017 1.032

10,023 9,911 9,922 11,100 10,242

1.021 1.044 1.056 1.017 1.040

10,287 9,961 9,988 11,239 10,741

1.017 .998 1.075 1.024 1.062

s Ratio: Deaths/Difference. Sources: Census of 1920, Descriptive Report, Appendix, Table 15. Vital Statistics, yearly volumes, 1906-1920. Statistical Summary, Table 58, for population by age in 1905.

1.094 1.014 .967 .992 .979 .971 .955 1,078 1.035 .053

PUBLIC HEALTH AND THE RISKS OF DEATH TABLE 35 TEST FOR ADEQUACY OF BIRTH REGISTRATION OF TAIWANESE

Births from 1906 to 1920, as Registered and as Estimated from Survivors Enumerated in Census of 1920 and Intervening Deaths Year of Birth

1920 1919 1918 1917 1916 1915 1914 1913 1912 1911 1910 1909 1908 1907 1906 1920 1919 1918 1917 1916 1915 1914 1913 1912 1911 1910 1909 1908 1907 1906

Survivors 1920

Registered Deaths Since Birth

Estimated Births (1) + (2)

Registered Births*

"Completeness" of Registered Births]

46,832 55,006 52,260 52,245 46,234 48,109 49,637 48,427 48,810 48,388 46,609 44,540 41,626 43,003 41,598

7,725 12,993 18,219 19,854 19,805 22,322 23,109 21,550 20,961 21,159 20,806 20,780 20,353 20,074 20,008

Males 54,557 67,999 70,479 72,099 66,039 70,431 72,746 69,977 69,771 69,547 67,415 65,320 61,979 63,077 61,606

54,064 70,349 71,764 73,430 65,837 70,855 72,755 70,170 69,990 69,558 68,042 65,666 61,993 63,290 62,120

.9910 1.0346 1.0182 1.0185 .9969 1.0060 1.0001 1.0028 1.0031 1.0002 1.0093 1.0053 1.0002 1.0034 1.0083

46,017 53,950 50,328 49,315 44,230 45,348 46,245 45,434 45,403 45,373 43,251 41,268 38,292 38,594 37,323

Females 5,815 51,832 10,346 64,296 16,028 66,356 18,099 67,414 18,658 62,888 21,372 66,720 22,279 68,524 21,281 66,715 21,090 66,493 20,805 66,178 20,262 63,513 20,187 61,455 19,737 58,029 19,981 58,575 19,824 57,147

51,410 66,358 67,701 68,984 62,768 66,814 68,695 66,797 66,632 66,100 64,099 61,620 57,807 58,466 56,987

.9919 1.0321 1.0203 1.0233 .9981 1.0014 1.0025 1.0012 1.0021 .9988 1.0092 1.0027 .9962 .9981 .9972

° In 1920, deaths and births have been taken only from January 1 to October 1. f Ratio: Registered Births/Estimated Births. Sources: Census of 1920, Descriptive Report, Appendix, Table 15. Vital Statistics, yearly volumes, 1906-1920.

PUBLIC HEALTH AND THE RISKS OF DEATH

the registration system were very slight. The same result is found for births of more recent years. With very few exceptions, these events—births, deaths, and pres­ ence at each census date—were nearly all recorded and were recorded accurately. The pattern of omissions is itself significant: whatever their source, they contained practically no age or sex bias. Therefore (with one reservation)19 we shall accept all published deaths and all population figures as complete and correct for the measurement of mortality. The differences would be trivial if they were distributed as adjustments over all the years of deaths; and the tests give no clue as to how such adjustment should be distributed. T h e Decline of Mortality

The simplest question that arises about the trend of Taiwanese mortality is also the easiest to answer: what happened to the prev­ alence of deaths in the population during colonial rule? Unfortu­ nately for judging this trend, no one can say what was the level of mortality in Taiwan when the Japanese first arrived. The Manchu government did not even know the number of its subjects, far less their behavior with respect to births and deaths. The colonial regime founded by the Japanese started a thorough system of vital statistics only after a decade of control. Consequently the very earliest usable record began at a time when life had already been made more secure. The trend of annual death rates after 1905 appears clearly in Table 36. It was not only downward, but was rapid in its fall. Over the same period there were scarcely any recorded instances to equal the speed of this decline. Supposing that there were over 40 deaths per thousand persons in 1895 and before, the yearly incidence must have dropped by more than one-half in the next fifty years. Thanks to improved drugs and newer approaches to the problems of health, such a feat has since this time become fairly easy in coun­ tries of high mortality. Crude death rates of 20 are no longer con­ sidered low. But subsequent advances in medicine should not be allowed to detract from this performance. With the resources and 19 The only reservation is in the case of infants. This requires special attention, and is discussed in the section on infant mortality. In reality, the ambiguity is a more fundamental one than any amount of "correction" can remedy. It arises from an inherent difficulty in making a precise definition for a "live" birth in a society where infants are delivered without skilled medical care and where the chances of surviving the first day of life are very poor. Where families regard mere discussion of deceased offspring as inauspicious, the prospects for developing pre­ cise measuring standards for infant mortality are poor.

PUBLIC HEALTH AND THE RISKS OF DEATH TABLE 36 CRUDE DEATH RATES OF TAIWANESE, 1906-1943» Years

Deaths per Thousand Persons

1906-10 1911-15 1916-20 1921-25 1926-30 1931-35 1936-40 1941-43

33.4 28.6 31.0 25.0 22.1 21.2 20.6 18.5

9 Unweighted averages of annual death rates for each year. Before 1915, deaths have been adjusted slightly on the basis of the internal pattern of infant mortality in later years, as explained in the text. Source; Statistical Summary, Table 64.

knowledge available in these earlier years it was a far more note­ worthy achievement than it now appears. Almost entirely through the reduction of mortality, the annual natural increase of Taiwanese rose to 25 per thousand, which even today is rarely found. The fall of mortality was not unbroken. It was interrupted in quick succession in 1915, 1918 and 1920 by waves of fatal illness over the island. The first occasion was accompanied by an unusual number of malaria deaths,20 when the program of malaria control was only in its second year. In 1918 and 1920, the Taiwanese were visited by a part of the world-wide influenza outbreaks at the end of World War I. In all three years the cause-of-death data, defective as they were, revealed exceptional fatalities due to malaria and re­ spiratory ailments respectively, but no serious recurrence of the old epidemic diseases. None of these setbacks quite matched the initial years of statis­ tical record. Since mortality in 1906 undoubtedly was already lower than it had been in the pre-Japanese period, one can say that for the island as a whole these later reversals were minor and transitory (see Fig. 21). Taiwanese Life Tables

The life table is a device that represents mortality in a special conventionalized way. Conceptually it is an instrument of precision, 20 And reportedly some civil disturbance, though such instances are always exag­ gerated and difficult to evaluate.

P U B L I C H E A L T H AND T H E RISKS OF

DEATH

Fig. 21. Crude Birth and Death Rates and Total Registered Births and Deaths of Taiwanese, 1906-1943.

for it allows one to speak in terms of probabilities of dying, which is what we mean here by the notion of mortality. Being conventional, it lends itself easily to comparison with data from other places independent of differences in age structure, even when the procedures of computation are slightly varied. Some caution, of course, must be 147

PUBLIC HEALTH AND THE RISKS OF DEATH

exercised in this application, for the life table is also a way of making poor mortality data look good. Since precise information about mor­ tality has not heretofore existed for a population consisting of Chi­ nese,21 tables prepared for Taiwanese under differing mortality con­ ditions should be of some additional interest from this standpoint alone. Life tables are very convenient as a way of summarizing agespecific risks of mortality for each sex in a specified short period. They can also be used to illustrate one aspect of the course of mor­ tality with the passage of time—the age incidence of change. There are more efficient ways of expressing secular change per se;22 where secular change is interrupted by fluctuations of short duration, the age incidence of deaths in a given period is probably the sort of measure to be preferred. In the case of Taiwan the life table is the most appropriate form in which to express the import of this particular type of death sta­ tistics. The data are peculiarly suited, as we have seen, to showing precisely the proportions of persons entering each year of age who died before the next birthday. This proportion is the figure con­ ventionally used to estimate the probability of dying in the given year of age. Having such data as these, one can prepare life tables directly and not by the roundabout and arbitrary procedures that ordinarily must be used.23 Actuarial practice has adopted strong graduation formulae to make life-table data appear more regular than they are. Presumably these procedures are grounded on assumptions that mortality risks are fundamentally a continuous function of age. On the basis of wide experience, this assumption seems reasonable enough. How­ ever, the justification for smoothing is based also on the assumption that much of the fluctuation in observed data represents chance variations in the risks of death at given ages;24 actuaries, understand21 Little attention has been given to two official tables that have been published. See Taiwan, Government-General, Research Bureau. Life Tables of Taiwan, 1926-1930 (First Issue) (1936), 160 pp.; Taiwan, Provincial Government, Department of Statistics. Life Tables of Taiwan, 1936-1940 (Second Issue) (1947). 22 A more satisfactory procedure for this specific purpose is to examine mortality of actual cohorts of persons over time, where age and time can be observed while varying together, and their interaction can be taken into account. It is planned to examine this aspect of mortality change in Taiwan at some later date. 23 Where S1 represents the number of survivors to exact age x, and D, the number of them who died in age x, qx (the probability of dying at that age) can be expressed directly as Dj/S,. The arbitrary conversion of an observed central death-rate, 111,, such as by assuming q,=2m,/ (2+mx), then becomes superfluous and can be by-passed. 24 See Knibbs, G. H. (1917), pp. 86-88.

PUBLIC HEALTH AND THE RISKS OF DEATH

ably, do not relish the thought that their work consists merely of manipulating errors of observation.25 Taiwan's data throw some light on the relative magnitude of the random and the error components of irregularity in age-specific mor­ tality rates. For example, one life table has been prepared for the period 1909-1911, in order to illustrate the risks of death at a time when they were still high. The mortality rates of this table (see Fig. 22) represent a small number of deaths, as do those of the suc­ ceeding tables, and have not been subjected to any graduation.26 The people in the lowest to the highest ages of any life table rep­ resent groups born over a long period of time, each of which might be expected to exhibit peculiarities of its own previous exposure to mortality. Nevertheless, the risks of death shown in Fig. 22 deviate only slightly from a smooth progression with age. The persons ex­ posed to death at each age were few; in view of their number, much larger variability would appear reasonable. Even if we were to as­ sume a complete lack of error in the data (throwing the whole burden of all deviations from continuity to the element of chance), the random component of these fluctuations is very small. A strong case for the notion of continuity can be made from this life table. But by the same token it appears that the error component of devia­ tion from continuity in ordinary data is the major portion. Perhaps not much general significance for the variability of "true" mortality rates by age can be drawn from these data. Very likely, however, smoothing devices as commonly used have as their major function the adjustment for errors of observation. Aside from errors of omis­ sion, these presumably arise because information by age is incorrect and because deaths as usually registered do not belong to the people to whom they are assigned.27 Taiwan's mortality data have not been widely known. Two official 26 There is in addition a possible third influence—the cohort effect, or effect of the date of birth. This is generally ignored. 26 Strictly speaking, they represent the weighted averages of three cohorts of persons entering each year of age, in 1909, 1910, and 1911 respectively. The mortality rates are in the simple form (D1-J-D2-I-Ds)/ (S1-J-S2-J-Ss) in each year of age. This averaging has the effect of reducing the variability of single-year values as compared with the life tables to be presented later. The effect of different weighting for different cohorts is ignored. 27 All this is not meant to imply that actuarial procedures are misguided or that smoothing is unwarranted when dealing with inferior data. The actual errors intro­ duced in most instances by graduation of mortality rates are probably not very great. It does seem important, however, to know how close an approximation to continuity accurate data can give, as a basis for inferring that the error component of deviations is not inconsiderable.

PUBLIC HEALTH AND THE RISKS OF DEATH

.180

MORTALITY, TAIWANESE

1909-1911 .1 6 0 -

.140-

.1 20-

.100-

MALE

.080-

.060-

.040-

FEMALE

.020-

0

IO

20

40

30

50

60

70

AGE OFFICE OF POPULATION RESEARCH, PRINCETON UNIVERSITY

Fig. 22. Mortality of Taiwanese, Life Table of 1910-1911. Values of q*, by Age, Male and Female.

life tables were compiled and published from them,28 which have likewise not become well known to Western demographers. Conse­ quently it is desirable to compare the conditions in Taiwan with 28 These two official tables were constructed by conventional procedures. The 1926193° life table, the earlier of the two, covered a period when the information on year o£ birth had been temporarily dropped from publication of vital statistics. The second table, for 1936-1940, covered a span when the census population could no longer be matched to registered deaths. Hence, the more precise procedure used here was denied to both of them.

PUBLIC HEALTH AND THE RISKS OF DEATH

SURVIVORS FROM LIVE BORN MALES

JAPAN, 1 9 3 5 - 3 6

TAIWAN

-TAIWAN, BRITISH INDIA, 1921-31·

O

IO

20

30

40

50

60

70

AGE OFFICE OF P O P U L A T I O N RESEARCH, PftlNCCTOH UWItfEBSiry

Fig. 23. Life Table Survivors of 10,000 Live Male Births, by Age (lx). Taiwan and Selected Other Countries.

other examples that are more familiar. This is done in Fig. 23, where life-table survivors of an arbitrary 10,000 male births have been plotted for Taiwan in 1906 and 1936-1940, Japan in 1926-1930, and British India in 1921-1931. Even near the turn of the century, to take the year 1906 as an example, mortality in Taiwan was no higher than that of British India many years later. The gross effects of mor­ tality in Taiwan had by 1940 come to a point closer to Japan during the 1920's than to the previous situation in Taiwan. The median age at death (the age at which mortality rates of the given period would have reduced a starting cohort by one-half) was nearly doubled between these two dates.

PUBLIC HEALTH AND THE RISKS OF DEATH

Such a comparison does not, however, reveal much about mortality in Taiwan itself. It shows next to nothing of the nature or extent of the decline. A better way to examine the shifts by age is to look at the changes in the risks of death. The probabilities of dying show these changes in each age independently of other ages. Graphically this profile of mortality appears for various dates in Fig. 24, plotted

MORTALITY RATES

200-r

200

MALES

FEMALES

160

-160

19 2 f - 2 2 1936 - 40

-120

80-

-80

40-

-40

20

40

60

70

0

AGE

20

40

60

70

AGE

OFFICE OF PQPULaTiQN RESEARCH. PRINCETON UNIVERSITY

Fig. 24. Mortality of Taiwanese, by Age (q*), Selected Life Tables. Males and Females.

on pivotal ages. The earliest point is the three-year period 1909-1911, expressing conditions that were greatly improved over those found in 1906.29 It represents the effects of selective sanitation works (i.e., in the larger cities), and of mass disease controls of the crudest sort. Epidemics had been suppressed, but the full-scale attack on malaria, 29 The procedures of preparing these life tables make them reflect the events of two calendar years. They are also based on the persons entering each age in the given year, some of whom did not die before the following year. The table for igo6, therefore, is composed mostly of deaths that occurred in 1906, plus some that did not take place until 1907. 1906 was the year of a severe earthquake in one Prefecture, and for this reason may have been less favorable than neighboring years.

PUBLIC HEALTH AND THE RISKS OF DEATH

for example, had not yet begun. Not long before, the Japanese had just succeeded in gaining undisputed control of the island outside of the central mountain region. By implication, the conditions in 1909-1911 thus give an indication of what is possible with the most rudimentary types of health measures. At that time most were still experimental, for the Japanese were just beginning to struggle with the problems of rationalizing agricultural practices, and had not had enough time to organize a concerted program against infection and disease. If mortality had been reduced easily and quickly, the gain was not a very secure one. The reversals in 1915, 1918, and 1920 are shown in Fig. 24 in their incidence by age; they again restored temporarily the unhealthful state that had prevailed before. (Influenza in 1920 struck early in the year, and affected principally those who had not passed a birthday since 1919; the age cohorts of 1919 were the ones that absorbed the force of the disease.) By contrast, when the steady effects of economic and health policies gradually became felt, mor­ tality risks (as exemplified by the period 1936-1940) were brought firmly down to much lower levels. Translated into figures for the mean expectation of life at various ages in each year, the risks of these years are summarized in Table 37. The pattern of the decline appears only when the mortality rates themselves are viewed in detail. From age 15 to age 50, most people pass through the most active portion of their lives. Adults in a country of relatively high mortality move from ages when death is rare through a stage of steadily increasing risks. By age 50 in Tai­ wan at least half of the members of each cohort of births were no longer alive. Adulthood is usually associated with two chief hazards of death. In a society like Taiwan, the working activities under­ taken by men and childbearing of women begin to bring people of both sexes into contact with greater risks of dying. The interval covers the entire normal childbearing period of women and most of the time of active employment of men. Deaths occurred in different patterns for men and women in these ages. Mortality actually started at slightly lower levels for women who were approaching adulthood than for men. It also advanced less rapidly with age. In years of moderate mortality this progression with age was nearly continuous; when there were extraordinary haz­ ards, they had the heaviest effect on women of younger ages and

PUBLIC HEALTH AND THE RISKS OF DEATH

TABLE 37 MEAN LIFE EXPECTANCY OF TAIWANESE AT DIFFERENT AGES

Selected Life Tables. Values of ex, in Years Age

1906

1909-11

0 1 5 10 30 50

27.7

33.2 35.9 33.5 20.4 12.0

32.4 38.1 40.0 36.7 22.6 13.2

0 1 5 10 30 50

29.0 34.0 38.8 27.2 25.5 15.5

35.6 40.7 44.7 41.9 29.0 17.4

1919

Males 27.7 33.4 36.2 33.4 20.8 12.5 Females 29.6 34.4 38.9 36.9 25.8 16.4

1921

1926-30

1936-40

34.5 41.6 44.0 40.8 26.2 15.0

38.8 45.9 47.5 43.6 28.1 16.0

41.1 47.6 49.5 45.6 29.7 16.5

38.6 45.0 49.2 46.4 31.9 19.6

43.1 49.4 52.5 48.8 33.5 20.3

45.7 51.5 54.6 50.8 34.8 20.6

Sources: Life tables of 1906, 1909-11, 1919 and 1921 were computed from the assembled data of Taiwan's censuses and vital registration. The table of 1926-30 is from Taiwan, Government-General (1936); that of 1936-40 is from Taiwan, Provincial Government (1947).

altered the entire pattern that usually prevailed (see Fig. 25).30 Thus enlarged from the mortality rates given in Fig. 24, the disparity be­ tween male and female age structure of adult mortality also appears more strongly. In spite of the differences in structure, male and female risks of death have both given the appearance of reacting in the same way to the changes in general conditions of health in the population. And this is very nearly true as far as secular decline in mortality was con­ cerned. But the exceptional risks (as in 1915, 1918, and 1920) affected women very differently from men. Their age incidence remained unchanged for men. For women, these risks were heavily concen­ trated at some of the later years of the childbearing period. The most typical pattern of sex selection of age-specific mortality is illustrated in the period 1936-1940 (see Fig. 26). The values are 80 Fig. 25 is based on age-specific death rates rather than on life-table probabilities. It makes no difference in the pattern of the results. Death rates are adopted for this illus­ tration because they show the events of the exact calendar years that are most appro­ priate for this comparison.

PUBLIC HEALTH AND THE RISKS OF DEATH

AGE SPECIFIC DEATH RATES, TAIWANESE

RATE PER

RATE PER

1000 PERSONS

1000 PERSONS

50

50 MALES

FEMALES 1906

40

40 1910

30

30 1918

1925

-1906

20

20

1940

1925 -

1940

15

20

30

40

50

AGE

15

20

IO

30 AGE

OFFtCE OF POPULATION RESEARCH, PRINCETON UNtVERSlTY

Fig. 25. Age-Specific Death Rates of Adults in Selected Years. Taiwanese Males and Females.

sex ratios of deaths at each age to an equal number of males and females beginning that age during the year in question.31 The usual sex balance of mortality risks in Taiwan was different from what prevailed in other recorded Asian populations—e.g., Japan and India. It is, of course, not surprising to see that it was not the same as the United States. But the nature of the difference also runs contrary to what might be expected. In particular, it is startling to find that females were normally less likely to die than males at all ages over 15 years. This critical age span, when women of a high-fertility so­ ciety are entering a period of frequent pregnancy and childbirth, is commonly thought to be likewise a stage of high mortality for them. The evidence as to fertility, which will appear in the following chap­ ter, shows that Taiwanese women did not avoid these risks by avoid31 Above age 4, the actual values plotted are sex ratios of q , or mortality rates over 5 x age intervals o£ 5 years each.

PUBLIC HEALTH AND THE RISKS OF DEATH SEX

RATIOS

MALE Cfr x FEMALE Cjj x

1.60 TAIWAN

1.40 1906-07-

1.20

1936-40 1919-20

1.00

.80

0

to

1.40

20

30

40

50

60

70

UNITED STATES, 1929-31

1.20

JAPAN, 1926-30

1.00

.80

0

INDIA, 1921-30

10

20

30

40

50

60

70

AGE O F F I C E OF P O P U L A T I O N R E S E A R C H , PRINCETON U N I V E R S I T Y

Fig. 26. Sex Selectivity of Mortality in Taiwan and Other Countries. Number of Male Deaths for Every Female Death, Among an Equal Number of Each Sex Entering Each Age.

ing childbirth. Fertility in Taiwan was consistently higher than for all of Japan, and yet in Japan women in childbearing ages were put at considerable disadvantage with men in their chances of survival.32 32 In both societies, the cessation of childbearing is marked, for example at age 40 and above, by a decided improvement in likelihood of survival of women in relation to

PUBLIC HEALTH AND THE RISKS OF DEATH

Though Taiwanese women in most years succeeded in warding off more risks of death than men in early adulthood, they did not fare so well during times of greatest sickness or near-epidemic. In both 1906 and 1920 the situation was reversed. In all ages other than infancy up to mid-adulthood, females experienced higher risks than males. Their greater susceptibility in these times is as puzzling as their comparative robustness in other years. It did not move them much closer to the position of Japanese females in relation to males, for the pattern of selectivity was still distinct from that of Japan. Regardless of the year of observation, young girls consistently suf­ fered greater risks than boys. In the first year of life, as is usually the case, males invariably had inferior average chances of surviving to their first birthday. For the next nine or more years they more than made up for this initial disadvantage. Without fail, there were more male than female survivors to age ten from an equal number of each who managed to enter the second year of life (see Table 38). This was the case for all cohorts of which the record is available from the first birthday to the tenth—-26 cohorts in all. Because of this, a higher proportion of boys than girls survived from birth to age ten. It is easier to reconcile lower rates of survival for young girls with known features of Chinese society than it is the advantage held by women over men at all higher ages. With a few exceptions, the dif­ ference in risks was consistent in every year of age. It often extended to age 12 or beyond, after which males generally remained at greater risk for the rest of life. Sex ratios of even numbers of one-year-olds surviving to age ten show a steady movement toward equalizing the risks of death in this interval with the passage of time. It is well known that male offspring are especially cherished in Chinese fam­ ilies, for they provide the basis for family continuity over several gen­ erations. The resulting preferential treatment in childhood for boys rather than girls is borne out by the evidence of Table 38, since sur­ vival of young girls elsewhere in the world is usually found to be superior to that of boys.33 men. So it is barely conceivable that unhealthful activities of men simply over-balanced these special risks of women. (Due to tuberculosis in Japan, male mortality almost out­ weighed that of females for a short span between age 20 and age 30, despite the fact that mortality for women was high.) It seems unlikely, however, on comparison with the pattern for India (Fig. 26). India was similarly an unhealthful place for men, but placed women at an even greater disadvantage. 33 A trace of this same phenomenon can be found in the early life tables for Japan. These are based, of course, on a particular period and not on actual cohorts. In the 1947 table for all Japan, however, mortality of males was higher at every age.

PUBLIC HEALTH AND THE RISKS OF DEATH TABLE 38 TAIWANESE SURVIVORS THROUGH EARLY CHILDHOOD (After Infancy) Actual Survivors to Exact Age 10 of Each Thousand Males and Females Who Passed their First Birthday, by Year of First Birthday Year

Males

Females

Male Survivors per 100 Females

1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934

812 822 819 825 826 820 819 813 807 812 810 814 825 843 853 865 856 852 862 863 870 870 874 870 866 873 870 870

769 781 785 792 790 784 777 777 773 780 778 787 801 822 834 846 840 834 841 846 854 854 858 855 848 861 858 858

106 105 104 104 104 105 105 105 104 104 104 103 103 102 102 102 102 102 102 102 102 102 102 102 102 101 101 101

Source: Based on Vital Statistics of each year. In most of the world, more deaths occur in infancy than at any other year of age. This is especially true of Asia. Deaths before com­ pleting the first year of life used to eliminate close to one-fifth of all births in Taiwan. In many countries this figure is believed to be

PUBLIC HEALTH AND THE RISKS OF DEATH

higher; cases of 25 per cent wastage of births in infancy have been known. Infant mortality is important for life tables also because it is difficult to measure. Most countries with high risks of death have been unable to keep adequate records of births, and infants who die after a few days or weeks in the world may not be accounted for either as births or deaths. The tests of completeness in registration and enumeration have demonstrated that in Taiwan nearly all chil­ dren whose births were registered were also either enumerated in subsequent censuses or recorded as deaths if they did not survive. The system of regulation that kept the local registers was severe as well as thorough. An entire community was held responsible, through its pao-chia organization, for failures to keep the registers in order. Some kinds of oversight might be successfully concealed, but if a child survived, its legal status as a person depended on hav­ ing a record with the police, which was initiated by entering the event of birth. Pressure brought to bear by neighbors, if not prudence on the part of the family itself, tended to promote compliance with these rules. Consequently, a child could not easily remain alive very long and escape being entered in the official records.34 There remains, however, a problem of unregistered non-survivors, who might have died so young that no official cognizance was ever taken of their existence or decease. There are three kinds of evidence that some non-conformity of this sort existed early in the registra­ tion period. First, birth rates (Fig. 21) appear to have risen slightly, which though possible seems unlikely. Second, the rates of infant mortality that one can derive from unadjusted registration figures seem lower than they must have actually been, in view of the decline in mortality at other ages.35 Finally, the sex ratio of registered live births was somewhat more masculine at the beginning than in later years. There is reason to suspect that some female infants might have been destroyed and not registered.36 34The government made provision for late registrations; there were very few, from which one can infer that almost all registrations that were made occurred soon after birth. 36 All figures for infant mortality before 1916 that are shown here have been adjusted slightly, by a procedure that is described below. Unadjusted, the data make infant mortality appear to have increased over the first several years of record. One is led to believe that the opposite was true, in view of the fall of mortality at other ages. 36It is not easy to determine the extent or even the general existence of early in­ fanticide in Taiwan. Missionaries and Western travellers often expressed indignation at the practice (see Pickering, Campbell, Davidson). Writings on both China and Taiwan have many references to it, mostly only gossip. Japanese officials took a strongly hostile attitude toward this disposal of infants, and apparently succeeded after a few years in

PUBLIC HEALTH AND THE RISKS OF DEATH

This idiosyncrasy of the registration system did not prevent our tests from demonstrating almost perfect correspondence among reg­ istered births, deaths, and children counted in the census. In other words, the omissions were exactly the same for births as for deaths. Internal evidence indicates that they were not permitted to continue very long. Infant deaths were also tabulated for small intervals within a year, from which probabilities of survival by month can be esti­ mated. The probabilities for the first month of life form a reason­ ably close regression on those for the second and third combined, for all years except 1915 and earlier. Between 1906 and 1915, the probabilities as recorded in the first month were higher than the values that the regression would imply. If we assume the regression of yearly observations after 1915 to be correct, and the probabilities in the second and third months of life to be accurate throughout, then the values for the first month reflect omissions before 1916 of babies who did not survive to one month of age. Following this assumption, all figures for births and infant deaths before 1916 that are used here have been adjusted upward slightly from their regis­ tered number to fit what regression values of these probabilities imply for infant mortality and births.37 Two procedures are often employed to measure infant mortality. The more appropriate one is to compile it as the proportion born who died in the first year of life, based on births occurring in a cal­ endar year and the deaths to this cohort of births before their first birthday. This is the measure that corresponds to the concept of lifetable mortality, and that will be used here except where otherwise specified. Ordinarily, however, deaths are not known by the year of birth of the decedents, but only by the calendar year of occurrence and by age. In this case it is customary to compute what we shall here call infant death rates, using simply the ratio of calendar-year births and calendar-year infant deaths. stamping it out. By 1906, when the vital statistics records can be relied upon in most matters, there was scarcely any trace of infanticide among Taiwanese. To set limits on the extent of the practice, we may assume that all male births were registered, that infanticide occurred only to females, and that the true sex ratio of live births was 105. Then (the sex ratio of registered births in early years being 109), in the first five years not more than 2 per cent of female babies could have been disposed of in this way. In any case, the sex ratio of registered births soon fell to 105, and stayed between log and 106 thereafter. 37 This is a conservative correction, since it entails the further assumption that these were the only omissions. The number of survivors to the end of the first year of life remains the same as before the adjustment.

PUBLIC HEALTH AND THE RISKS OF DEATH TABLE 39 INFANT MORTALITY OF TAIWANESE*

Deaths Per 1,000 Live Births in Each Year Males

Year

Death Rate (Calendar-Year Deaths)

Females

Mortality Rate (Actual Deaths)

Death Rate (Calendar-Year Deaths)

Mortality Rate (Actual Deaths)

1906 1907 1908 1909 1910

192 174 190 182 167

182 175 188 183 168

173 161 173 162 144

165 162 172 163 144

1911 1912 1913 1914 1915

176 164 177 180 203

171 169 178 187 202

144 139 149 161 183

141 143 152 167 180

1916 1917 1918 1919 1920

193 173 202 186 190

179 182 199 196 178

161 148 172 160 161

150 155 173 167 149

1921 1922 1923 1924 1925

173 184 164 190 186

175 180 171 196 177

144 155 133 159 158

147 150 141 164 150

1926 1927 1928 1929 1930

171 175 181 171 163

1931 1932 1933 1934 1935

177 167 161 167 160

1936 1937 1938

159 157 159

— — — —

—

139 145 152 141 139

— — — —

—

169 163 163

146 140 137 143 133

143 139 133

159 159 155

131 132 132

133 134 128

— —

161

—

PUBLIC HEALTH AND THE RISKS OF DEATH

TABLE 39 (continued) Males

Year

Death Rate (Calendar-Year Deaths)

Females

Mortality Rate (Actual Deaths)

Death Rate ( Calendar-Year Deaths)

Mortality Rate (Actual Deaths)

1939 1940

152 148

156 136

126 122

129 114

1941 1942 1943

131 136 134

133

114 116 116

116

—

— —

0 Registered deaths have been adjusted slightly before 1916, based on a regression of mortality in the first month of life on mortality in the second and third months of life in 1916 and all subsequent years. Births have been adjusted accordingly. Data for actual mortality rates were not published for certain years. Sources: Vital Statistics, individual years; Statistical Summary, Tables 64 and 89.

We must rely on death rates of this latter sort to examine the internal composition of infant mortality, for deaths by month-of-age were published only in this form. This approximation measures in­ fant mortality with nearly the same yearly values as did the actual rates (see Table 39). It is worth while to tolerate some disagreement between the two in order to divide mortality of the first year into smaller intervals. The progression of these death rates through time is plotted in Fig. 27. Infant mortality on the whole did not decline very sharply. (There will be occasion shortly to compare it with other ages.) If our guess about the defect in registration is correct, then two con­ clusions necessarily follow. First, the risks of death after the first month of age are documented fully, without omissions. Second, and more important, mortality in this span—the so-called postneonatal mortality—dropped no more than all infant mortality. Neonatal mortality then changed just as sharply as that of infants who survived the first month.88 Such a pattern of decline in infant deaths is contrary to what is commonly taken for granted. The usual suppositions about infant mortality in the process of decline have been made a priori, for 38 In addition, according to our guess above, if additional events (births and deaths) ought to be added to the earliest years to make the figures absolutely complete, they must be added to the neonatal period, making mortality in the first month still higher at the beginning. This could only have the effect of indicating a more pronounced drop in neonatal mortality. There is almost no way of allocating extra deaths that could reverse the pattern shown in Fig. 27.

PUBLIC HEALTH AND THE RISKS OF DEATH INFANT DEATH RATES, TAIWANESE MALES

.10 MVtONTH .08 .0 6 Six

.04

.02

.01 1905

3 - 5 MO,

I 1910

I 1915

I (920

I 1925

1 (930

I 1935

I 1940

OFFICE OF POPULATION RESEARCH, PRINCETON UNIVERSITY

Fig. 27. Fall of Death Rates of Infants by Month of Age. Taiwanese Males, 1906-1940.

actual evidence is rare.39 Neonatal mortality is thought to be resist­ ant to measures that bring infant mortality under control and thus to most general health measures. Part of the reason may lie in par­ ticular environmental influences such as standards of nutrition in a population, the climatic conditions, and the customs governing child­ birth and infant care. Part also may be linked with maternal health, which could exert a pre-natal as well as post-natal effect on the chances of live-born babies to survive. It is curious that the distinc­ tive behavior of neonatal mortality in Taiwan (as compared with Europe and America) should coincide with the distinctive low mor­ tality of adult women in childbearing ages. Whether or not one has influenced the other is impossible to say without actual information on the matter. The medical aspects of reproduction in such an Asian 39 The evidence also has come from European and American countries. Unfortunately it has been subject to a large sampling bias, inasmuch as it represents those countries whose progress toward low mortality has occurred within a particular cultural order, of which the individual countries are but variants.

PUBLIC HEALTH AND THE RISKS OF DEATH

cultural order are so little known in the West that the question can­ not be pursued further except by speculation. Decline of Mortality by Age

The appearance that the data above give to infant mortality may lead us to keep certain reservations about their accuracy. No such doubt can exist about the general validity of mortality registered at age one and above. The tests reported in Table 34 and Table 35 present Taiwan's statistical system as a paragon even among those of the technologically advanced nations in the world. The trends they show in mortality of later life cannot but be accepted. For what they are worth, our best estimates of mortality in in­ fancy suggest that it did not decline sharply. On surer ground, it is certain that the chances of dying in later years of childhood and early adulthood fell faster. Fig. 28 displays graphically the basic data on rate of decline in mortality (chances of dying) at selected ages of males entering those ages each year.40 In the ages taken here as examples—1, 2, 3, 4, 10 and 25—older children and young adults gained the most in chances of survival. Among children, in general the older, the more the improvement. Every higher age gained more ground than infants. Part of this difference may be due to the nature of changes made through colonial administration in the external risks confronting persons of each age. But doubtless a certain part is also due to the effect of carrying improved health from one age to the next by the same individuals—the "cohort effect." Yearly fluctuations are large enough, however, that no special patterns of either type are to be found without complicated manipulations of the data. To find the risks declining more or less together at different ages does not imply that the number of years of life saved in the popu­ lation was the same with advancing age. The saving of life in in­ fancy was greater, despite the more moderate drop in the risks than at any other age. This fact holds true regardless of whether one adopts this period-by-period approach or follows the mortality of successive actual cohorts. 40 The earlier values of the mortality rates at higher ages are connected with broken lines because the observations are not spaced at every year. Mortality rates for every year have been computed only for persons born after 1905.

PUBLIC HEALTH AND THE RISKS OF DEATH

PROBABILITY OF DYING ,TAIWANESE MALES

UNOER AGE I

.20

.08

I YEAR OF AGE

.06 -2 YEARS ,04 3 YEARS

.02

25 YEARS 20 YRS

.008

IO YEARS

.006

.004

.002

.001 1905

1910

1915

1920

1925

1930

1940

OFFiCE Of POPULATION RESEARCH. PRINCETON UNtVESSlTV

'

Fig. 28. Decline of Mortality by Age, 1906-1941. Taiwanese Males.

Sectional Differences in Mortality

It would be strange if for people of each age the risks of death were the same everywhere in Taiwan. It happens that most of our information can be obtained only for the island taken as a whole. The measures have been designed to take advantage of this and to present a detailed view of the total population at some sacrifice of knowledge about its parts. Moreover, with a total population that

PUBLIC HEALTH AND THE RISKS OF DEATH

was so small, any reduction in the number of cases tends to make sensitive measurements unreliable. The livelihood of Taiwanese farmers brought them into close and continual contact with their physical surroundings. Such a people cannot avoid some influence of climate and environment on its chances of survival. For a place so small there were marked differ­ ences in seasons and rainfall over the island. Whether or not related to this factor, from time to time local peculiarities in the incidence of disease have often been noted. In pre-Japanese times, isolation of small localities kept much of the communicable disease from being transmitted across the island. Even epidemics were observed that stayed confined to towns and apparently did not become general. The Japanese destroyed this barrier by an efficient network of communications, but maintained special medical inspections on the railroads and at check points on routes of heavy travel. They also imposed port restrictions to pre­ vent disease from being imported from abroad. As the site of the most modern cities and the favorite residence of Japanese, Taipei Prefecture was one of the safest places in which to live. It was also more liberally supplied with medical facilities than any other Prefecture, because it was Taiwan's center for training and research in tropical medicine. It was relatively free from malaria, and contained the most elaborate sanitary facilities in the island. Other sections were less distinguished, unless by exceptionally heavy incidence of sickness. Taichung was the most afflicted with malaria. The three southern Prefectures were all said to have an unusual amount of influenza and other respiratory disease. The smaller settle­ ments along the eastern coast started out with health conditions far inferior to the rest, though they were gradually improved. Whatever the causes, there were wide disparities in age-specific mortality among the various Prefectures as late as 1925 (see Fig. 29, for males only). For some unknown reason Hsinchu was the most favored part of the island, followed by Taipei. The range from high­ est to lowest at this date was almost as great as between the begin­ ning and the end of the Japanese period. It is interesting also that the differences were consistent at all ages, with almost no shifts in the standing of Prefectures at any age. At a time when there was very little migration between Prefectures, these life table values may be regarded as a good representation of actual differences. Expressed as the mean expectation of life at selected ages (Table

P U B L I C H E A L T H AND T H E RISKS OF

DEATH

Fig. 29. Mortality of Taiwanese Males in Each Age, by Prefecture. Values of q,, Life Tables of 1924-1925.

167

PUBLIC HEALTH AND THE RISKS OF DEATH

TABLE 40 MEAN LIFE EXPECTANCY OF TAIWANESE AT DIFFERENT AGES

Prefectural Life Tables of 1924-1925. Values of ex, in Years P R E F E C T U H E

Age

Taipei

Hsinchu

0 1 5 10 30 50

38.3 45.2 47.5 43.6 28.0 15.4

43.1 50.2 51.5 47.4 31.1 17.1

0 1 5 10 30 50

42.2 48.3 52.8 49.2 33.7 20.1

46.6 52.3 55.1 51.4 35.4 20.4

Taichung

Tainan

Kaohsiung

30.4 37.1 39.7 36.5 22.3 12.4

31.7 39.2 41.8 38.3 23.8 13.2

34.7 40.8 45.0 41.9 28.3 17.0

34.5 41.4 45.8 42.7 28.8 17.2

Males 35.9 42.8 45.7 42.0 26.0 13.9 Females 40.7 46.9 52.0 48.5 32.6 19.1

Sources: Computed from census and vital registration data of 1924 and 1925.

40), the effects of mortality were as heavy in Tainan as for the total island in igog-igii. On the other hand, the expectancy was even higher at all ages in Hsinchu than for all of Taiwan in 1936-1940, and much higher in the ages of childhood (cf. Table 37, above). Mortality according to these data tended to align itself with the axis of the island—low in the north, high in the south, and inter­ mediate in the central part. For a land so small, and a population of so few, sectional differences of this magnitude are indeed sur­ prising. (The two small eastern Prefectures have been omitted, not having enough people to make a meaningful life table.) This pattern of dissimilarity was of long standing. Turning to cruder indices, we can see that the same order of ranking existed throughout the registration period (see Table 41). No matter what index of mortality is used, the grouping remains almost unchanged. Here also it is evident that the places of higher mortality were making the faster progress toward control over death than the ones that were already low. This appears as well in the fact that the rela-

PUBLIC HEALTH AND THE RISKS OF DEATH TABLE 41 TAIWAN. DIFFERENCES IN MORTALITY BY PREFECTURE, IN SELECTED YEARS

Crude Death Rates, Infant Death Rates, and Deaths Per Thousand Females Aged 15 to 49 f PREFECTURES'*

Year

Taipei

1906 1915 1925 1930 1935

28 25 23 22 22

1906 1915 1925 1930 1935

162 157 157 154

1906 1915 1925 1930 1935

13 11 9 8 8

—

Taitung ir P'enghu Hsinchu Taichung Tainan Kaohsiung Hualien Islands

22 20 19 17 19 —

Crude 37 30 24 21 21

Death Rates 42 41 36 34 29 29 23 24 21 21 Infant Death Rates —

—

—

35 30 31 21 21 —

138 148 176 192 191 149 139 164 191 180 130 153 152 161 179 139 160 146 151 155 Deaths per Thousand Women Aged 15-49 10 15 24 29 26 9 13 20 16 20 7 9 15 14 14 6 7 9 10 8 7 9 8 9 8

37 26 26 22 19 —

204 t 172 131 15 12 _+ +

9 5

Total Island

35 29 25 21 21 192 174 167 156 149 20 15 11 8 8

f All rates are averaged over three years around the census year. Those for 1906 represent one year only. Infant death rates have been omitted for 1906, as being unreliable. 0 Prefectural divisions before 1920 have been matched approximately to the subse­ quent boundaries. 1 Included in Kaohsiung Prefecture.

tive dispersion of death rates was continually being reduced along with the death rates themselves. As expressed in terms of the relative variation among Prefec­ tures in Table 42, the gains in survival were being spread more evenly over the various portions of the island as time went on. And part of the reduction in mortality for the total must have taken the form merely of levelling some of the initial inequities in health

PUBLIC HEALTH AND THE RISKS OF DEATH TABLE 42 RELATIVE VARIATION OF MORTALITY AMONG PREFECTURES IN SELECTED YEARS

CoefFicieiits of Variation for the Death Rates of Each Year in Table 41 Y ear

Crude Death Rates

Infant Death Rates

1906 1915 1925 1930 1935

19.2 17.7 15.2 9.6 5.1

13.2 10.8 9.3 6.3

—

Deaths/ Women 15-49

36.0 27.9 27.3 15.2 16.6

Source: Computed from data of Table 41.

conditions. Much headway was made in this process after 1925, when the life tables above revealed such large gaps in the degree of con­ trol over sickness and death. It must be kept in mind when studying mortality here and in other areas of rapid decline that initial achieve­ ment is likely not to extend to many in the population, and that later progress may be no more than bringing more types of people nearer to the same point as the ones who were privileged in the beginning. Taiwan formerly had a well-earned reputation as a spot of capri­ cious and unchecked risks to life among its inhabitants. Its position athwart trading routes of the last century made it a likely place to introduce epidemic disease from the rest of Asia. As it became drawn into contact with the outside world and its wider sources of infec­ tion, the state of medical knowledge among its people did not keep pace. The former government lacked the capacity and the incentive to adopt measures that might alleviate or forestall such sickness. Taiwanese were therefore often subjected to intense attacks of these diseases, some of them becoming general over the island. In fifty years of administration the Japanese quickly eliminated epidemics. They succeeded in gradually making the other sources of fatal illness and ill health less prevalent, and thus postponed the prospect of death for Taiwanese. While some characteristics of the decline in mortality that they induced followed familiar patterns, it also carried certain distinctive traits of the society in which it occurred. Special features like the apparent vitality of adult women

PUBLIC HEALTH AND THE RISKS OF DEATH

and the importance of neonatal mortality in the decline of total infant mortality we have established with repeated observations. As far as present knowledge is equipped to deal with them, these pe­ culiarities must remain inexplicable. The means to this accomplishment were inexpensive and rudi­ mentary. Aside from probably raising levels of living, the Japanese developed a concerted program against an important group of dangers to health. They attacked those pestilent conditions that can be reached through sanitation and slight reorganizations on a mass scale of practices that were deeply rooted in custom. More thorough measures, which were also more burdensome, were avoided. Their health program relied on administrative efficiency and energy in execution to make up for what it lacked in elegance of equipment and technique. It is significant to find that such an approach could achieve the results that it did. In many other areas of high mortality the desire to reduce the risks of death is prompting steps in the same direction, and Taiwan's experience ought to prove instructive for such efforts. It is only a modest sample of what might be done in a "backward" region with effective administration—without necessarily incurring the great expense often thought to be associated with high standards of public health. Recent progress in research and development of drugs can make success more certain than before. Future strides in medical knowledge can enhance it still further. The greatest limitation is that health promotion among people with high mortality requires constant vigilance. Default threatens such a program with collapse, for medical procedures now known are still not automatic, nor are they foolproof. Japanese measures to control disease were primarily suppressive. They did not eradi­ cate malaria or remove the potential sources of cholera infection. Failure to appreciate these latent dangers led to severe outbreaks of cholera and malaria after the island was taken over by forces from the mainland in 1945. Since then the standards of administration, which had been allowed to deteriorate, have been raised again; with the improved technical knowledge at its disposal, the Nationalist government may be able to keep mortality at levels even below those of the Japanese period. The example of Japanese health policies, however, remains a useful lesson in what energetic measures can accomplish without the benefit of these superior medical facilities.

6,121 5,876 5,528 5,043 4,524 4,014 3,550 3,140 2,744 2,324 1,884 1,434 996 595 281 125 38

6,217 6,009 5,691 5,212 4,617 4,015 3,398 2,805 2,236 1/712 1,224 798 443 217 80 26 5

6,694 6,554 6,329 6,020 5,634 5,194 4,705 4,154 3,590 2,957 2,311 1,689 1,087 616 288 95 19

6,974 6,745 6,612 6,419 6,122 5,816 5,446 5,074 4,696 4,305 3,884 3,366 2,766 2,107 1,413 745 330 99

7,062

7,445

7,154 7,258 7,038 7,150 6,860 6,975 6,592 6,724 6,268 6,440 5,884 6,114 5,423 5,749 4,892 5,377 4,290 4,986 3,637 4,534 2,938 3,988 2,192 3,315 1,472 2,553 845 1,750 387 1,007 128 446 26 134

7,337

7,319 7,216 7,048 6,809 6,530 6,201 5,820 5,234 4,766 4,101 3,347 2,546 1,738 1,010 465 153 30

7,495

7,453 7,354 7,199 6,988 6,754 6,479 6,156 5,816 5,443 4,990 4,424 3,734 2,926 2,005 1,126 461 114

7,627

10,000 10,000 8,444 8,706 7,959 8,169 7,713 7,876 7,578 7,717

10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90

5

0 1 2 3 4

Males Females Age

1936-1940

6,337 6,159 6,811 6,717 6,189 5,945 10 6,139 5,980 6,624 6,543 5,949 5,680 15 5,872 5,704 6,379 6,273 5,646 5,353 20 5,344 5,477 5,886 5,986 5,222 4,948 25 4,938 4,954 5,486 5,491 4,707 4,514 30 4,355 4,511 4,924 5,058 4,123 4,061 35 3,725 4,061 4,301 4,605 3,479 3,615 40 3,063 3,620 3,625 4,183 2,841 3,193 45 2,413 3,201 3,739 2,944 2,217 2,753 50 2,696 1,812 3,227 2,278 2,285 1,642 55 1,260 2,150 1,658 2,674 1,153 1,819 60 812 1,586 1,139 2,092 735 1,344 65 456 1,056 697 1,498 892 414 70 573 212 931 368 502 191 75 244 80 477 137 73 232 80 108 26 212 45 103 24 85 32 5 9 64 31 5 90 Sources· Life tables of 1906 1909-11, 1915, 1919 and 1921 were computed from the assembled data of Taiwans censuses and vital registration! The table of 1926-30 is from Taiwan, Government-General (1936); that of 1936-40 is from Taiwan, Provmcial Government (1947).

6,607

6,613

7,072

7,125

6,498

5

6,668

10,000 10,000 8,257 8,546 7,775 8,004 7,551 7,705 7,420 7,541

10,000 10,000 8,080 8,377 7,590 7,811 7,290 7,450 7,097 7,209

10,000 10,000 8,037 8,328 7 422 7,611 7,061 7,158 6,802 6,838

10,000 10,000 7,980 8,200 7,363 7,472 7,045 7,033 6,829 6,750 6,683 6,562

10,000 10,000 8,261 8,510 7,728 7,892 7,446 7,491 7,253 7,232

10,000 10,000 8,078 8,270 7,458 7,552 7,100 7,061 6,841 6,721

0 1 2 3 4

1926-1930

Males Females

1921

Males Females Males Females

1919

Males .Females

1915

Males .Females

1909-1911

Age Males Females

1906

TAIWANESE LIFE TABLES Number of Survivors to 10,000 Live Births (Ix), Selected Years, by Year in Which Persons Entered Each Age

CHAPTER VI. APPENDIX

CHAPTER VII

The Setting for Family Life No MATTER what part of Taiwanese life we have investigated, it has not been possible to proceed very far without encountering evi­ dences of familial pressures and sentiments. This has been true of the organization of farming, the "inheritance" of skills, the reluctance to migrate, certain special patterns of mortality, and the attainment of peace and order by the Japanese. Yet kinship is something that people usually consider as a private matter (and moreover take for granted without inquiry as it exists among themselves), and so it is not easy to increase our knowledge of it. The aspect of kinship that makes it so important is the framework of obligations and expectations that it imposes on people's activities in related spheres.1 Like many forces affecting behavior, the Chi­ nese family system is an intangible thing, and must be approached by indirection—for example, by construing some of the basic moral precepts that the society has received from the past, which is the usual procedure, or, as we shall essay in this chapter, by examining some of the patterns that people display in their familial roles. The former of these devices has been singularly appropriate, for Chinese have one of the richest accumulations of expressions about their sentiments in family matters that ever existed; but it has left us with a one-sided and sometimes unclear view of the subject, largely because it has been followed with little reference to the latter. While neither of these alternatives gives a perfect representation of the family system in actual operation, the advantage here lies in using Taiwan's census data to illuminate a few points on which accurate information has previously been lacking. The Size of Household Units Japanese authorities showed that they were aware of the family's practical significance to the people, for in Taiwan they made it the basis of their census and vital registration statistics as well as of their regular civil administration. In each census the blank forms that were distributed to be filled in with personal information were all made out with the household as the unit of enumeration. The head ι For thorough discussion of the family's role in Chinese social organization, see Lang, O. (1946) and Levy, M. J„ Jr. (1950).

THE SETTING FOR FAMILY LIFE

of a household was the one responsible for entering the correct facts regarding each member. The Japanese owed much of their creditable success in census enumeration in Taiwan to the way they were able to secure conformity to these standards by household heads. Once they were in possession of authority to do this, the family system did the rest.2 The household is not identical with a family unit. It includes some persons living with a family group who are not related by birth or marriage to the others, and it leaves out all who are tempo­ rarily or permanently absent at the date of the census. But the house­ hold, properly defined, is the unit that counts in Chinese society. It is the concrete group in which joint decisions are made, and in which the obligations of the family system operate most directly. Also, it is the foremost, for some people the only, unit of participa­ tion in the society at large.3 Provided that sufficient information is given about its members, the household is probably a more useful basis for enumeration than kinship, strictly defined. 2 The high caliber of their census system did not owe everything to this expedient, however. It had been a common feature of the ancient registrations of people in China, where in pre-modern times it had not given such gratifying results. Japanese adapted the lessons of experience in these matters by Western nations, and added some innova­ tions of their own. They worked out sets of definitions which provided meaningful categories and consistent groupings of the population (as well, of course, as some that did not), and they developed an effective administration to carry out the work of enumeration. In such a society the most modern innovations in census technique are useless by themselves. Similarly, the family system without special statistical standards may be a poor instrument for recording information about its members—there is every en­ couragement to evasion unless this is forestalled. Systematic omissions might be more serious in household statistics than in any other type of tabulation, for they would probably be concentrated in certain types of household members whom there was some incentive to conceal—the biases would be strongest in a table that showed household membership. In order to control the responses in these household schedules, the census rules de­ fined an ordinary household as those persons actually present who had residence, family budget, or both, in common. People, or branches of a family, who did not meet these simple but effective requirements were enumerated as separate households; special types of living arrangements ("quasi-households," q.v. below) were assigned to another category. These and other rules of procedure were maintained with only minor changes of wording in all of the seven censuses of Taiwan. They were reported fully in the explanatory section or appendix of the main descriptive volumes for 1905, 1915, 1920 and 1930. s i n Taiwan there was another basis for organizing family solidarity, the so-called "clan" or tsu. It was very prominent, especially in pre-Japanese times when it was the basis for setting in motion insurrections and often the means for restoring peace and order. But the clan was never a family in the sense of a common unit of participation in the society. It was a loose alliance of affiliated families rather than the unit in which everyday matters were conducted. For the distinction between the two, and a discussion of the tsu as it appeared in South China, see Hu Hsien-chin (1948).

THE SETTING FOR FAMILY LIFE

Chinese cultural norms set no limit to the number of family rela­ tives that it is desirable to maintain in a single household, but at­ tach great merit to holding together many branches of a family in one unit. Though Chinese are thus inclined to feel that the bigger the family the better—there seem to be no general class or status differences in the ideal size that they should strive for—so many de­ grees of relationship are eligible for inclusion that other considera­ tions usually enter and help to settle the question. In practice, the real limitations tend to be how strongly a particular family feels it can enhance its prestige through adherence to tradition, and how well it can afford financially to do so. TABLE 43 SIZE OF HOUSEHOLDS IN TAIWAN IN SUCCESSIVE CENSUSES

Average Persons per Household, All Types of Household, All Ethnic Groups, 1905 to 1940 1905 1915 1920 1925 1930 1935 1940

5.2 5.3 5.3 5.5 5.7 5.8 6.0

Sources: 1935 Census, Statistical Tables, Descriptive Section, p. 30. 1940 Census, Summary Returns, published in the Gazette of the Govern­ ment-General of Taiwan, April 22, 1941.

Despite the considerable emphasis on their size and cohesion, these family units actually are not exceptionally large. From the total numbers of households and of people, published with the regu­ lar report of each census of Taiwan, we can first establish a few main features of the case to guide the more detailed discussion that is to follow. Expressed as the average number of persons to a house­ hold, these figures show that the size of households (for everyone, mixing together all nationality groups and all types of household arrangements) started out in 1905 at 5.2, and rose steadily during the census period (see Table 43). The increase was not especially rapid, but it was unbroken and marked by acceleration with the passage of time. Correlated with what is already known about the growth of population in Taiwan, this steady movement implies that

THE SETTING FOR FAMILY LIFE

households were not keeping up with the increase of people, and that family groups were therefore undergoing some change in their composition. Though some aspects of this change can be shown more satisfactorily below, the process is most easily seen in these figures, since they cover the complete span of the census administration. As usual, the published results of these census materials did not do justice to the care that went into their collection. Household tables were not always shown in useful form, and the figures of Table 43 do not convey very much even about the size of families in Taiwan. Obviously in order to draw reliable conclusions about families of Taiwanese it must be possible to separate them as a group in the statistics, and to make sure that the statistics employed are a fair representation of family groups. For exactly this purpose, the Japanese treated one kind of household as standard, the so-called "ordinary" household, to cover the "normal" types of living ar­ rangements. People in certain special categories were excluded; resi­ dents of these non-family households, or "quasi-households," will be omitted from most of the tables that follow.4 Information about quasi-households is sketchy and hard to find. The size of these establishments is unknown for separate nationality groups, perhaps through a deliberate effort to conceal the military forces (though they would nearly all have been classified as Jap­ anese anyway, according to the status of the person in charge). The figures can be pieced together for the entire population, and appear in Tables 44 and 45. Quasi-households were a very small part of the total, but accounted for a much larger portion of non-Taiwanese. Only one person in a hundred of all nationality groups lived in such circumstances; 6.6 per cent of Japanese and 20 per cent of Mainland Chinese lived in quasi-households in 1930, but relatively few of the Taiwanese were found outside of ordinary households. Since Tai­ wanese are the main source of interest in the problem, the few liv­ ing in quasi-households are unimportant, and for the remainder of this chapter the term "household" will be used to refer to "ordinary" households only. On the average, ordinary households of Taiwanese contained be* They comprise people counted in hospitals, prisons, schools, military barracks, police and other special dormitories, and hotels. Although they must be removed before the household data can be said to represent patterns of familial life, the line drawn between the two types of household is not entirely rigorous. Many of these people, especially the ones in hotels and schools, were only temporarily absent from their homes, and their family affiliations may have continued undiminished.

THE SETTING FOR FAMILY LIFE TABLE 44 PER CENT OF TAIWAN'S POPULATION LIVING IN "ORDINARY" HOUSEHOLDS, 1905-1930

Both Sexes

Year

Persons Living in Ordinary Households

Persons Living in "QuasiHouseholds"

Per Cent Living in Ordinary Households

1905 1915 1920 1930

All Nationality Groups 18,491 3,021,260 3,461,282 18,640 30,905 3,624,403 45,567 4,546,970

99.4 99.5 99.2 99.0

1920 1930

Taiwanese 3,454,956 4,293,234

11,551 20,447

99.7 99.5

12,404 14,967

92.4 93.4

Other Nationalities* 17,585 6,950 40,422 10,153

71.7 79.9

Japanese 1920 1930 1920 1930

151,862 213,314

* "Other nationalities" consists almost entirely of Mainland Chinese. Sources: Census of 1905, General Volume of Statistical Tables, Table 114. Census of 1915, Statistical Tables, Table 109 and 116. Census of 1920, Descriptive Report with Appendix, Appendix Table 149, 164, 165 and 169. Census of 1930, Statistical Tables, Total Island, Table 5.

tween five and six persons during that portion of the census period when they can be observed separately. Because Taiwanese made up almost the entire population and because most of them lived in family groups, the initial approximation of Table 43 was fairly close to the actual size of their households. Differences begin to emerge when other nationality groups are listed by themselves; the house­ holds of both Japanese and Mainland Chinese on the average were smaller by a considerable margin than those of the Taiwanese, and remained that way (see Table 45). Not only did Taiwanese follow the ideal of large households more closely than the -other population groups in Taiwan, but they dis­ played a different pattern in approaching it. Distributed according to the number of persons in each, only the households of Taiwanese

THE SETTING FOR FAMILY LIFE TABLE 45 SIZE OF HOUSEHOLDS, BY TYPE AND NATIONALITY OF HEAD

Average Persons Per Household, "Ordinary" Households and "Quasi-Households" All Nationality Groups, Taiwanese, Japanese and Other Nationalities 1905 to 1930 Year

Ordinary Households

QuasiHouseholds

1905 1915 1920 1930

All Nationality Groups 5.19 5.31 5.30 5.64

6.2 12.4 20.6 17.1

1920 1930

Taiwanese 5.45 5.82

1920 1930

Japanese 3.37 3.74

1920 1930

Other Nationalities* 3.79 3.84

* "Other nationalities" consists almost entirely of Mainland Chinese. Source: From data of preceding table. Certain army and navy barracks were in­ cluded as quasi-housenolds in the Census of 1920 for the first time, accounting for much of the increase over 1915 and 1905 in the average size of quasi-households shown above.

were in a pattern that reveals a predominance of families with two or more generations, families of the sort that must be required to maintain a high rate of natural increase. Their households were found most commonly with four and five members, more rarely of smaller sizes. The process by which the average size of Taiwanese households rose from 1905 on was not a completely general shift toward larger units, but rather one of forming relatively fewer small households (with the exception of people living alone) and raising the proportion of very large ones (see Fig. 30). The sizes most often achieved remained very nearly the same as before, and the distribu­ tion simply became more skewed. (Expressed in terms of the num­ bers of persons involved, the effect of this change would, of course,

THE SETTING FOR FAMILY LIFE

SIZE OF HOUSEHOLDS

•1905

16

.*'1920 TAIWANESE

12 1930 θ

4

0 12 8 . 6 % )

PERSONS PER HOUSEHOLD

16 1930 12 TAIWANESE

8

4

O

OTHER NATIONALITIES

2

6

8

O

16-20 21-25 26

PERSONS PER HOUSEHOLD OFFICE OF POPULATION RESEARCH. PRINCETON UNIVERSITY

Fig. 30. Size of Households in Taiwan at Various Dates. Households Distributed by the Number of Persons in Each. Taiwanese, Japanese and Others.

THE SETTING FOR FAMILY LIFE

be multiplied in households of larger size, since they were larger by virtue of containing many people.) A great many Japanese and Chinese nationals staying in Taiwan obviously were separated from close family ties—for example, those who lived alone, and perhaps some of those in households of two or three persons, not to mention the ones who stayed in quasi-households. Few in either group maintained households in Taiwan as large as those that were usual among their countrymen at home. The Japanese gradually began to establish families in Taiwan, and to take up solitary living somewhat less frequently than at the be­ ginning of their rule. In contrast, an extremely high proportion (29 per cent) of Chinese from the mainland by 1930 were still living alone in households of single persons, a result of the transient and impermanent status that they occupied in the island. It is interest­ ing that in both of these minority groups, households of more than five persons, though not very numerous, had a distribution less unlike those of the Taiwanese. Apparently among the more settled and permanent non-Taiwanese in the population the ideals of the desirable family size were similar to those of the Taiwanese; but this attitude seems to have been more widely shared by Taiwanese, who as a group came closer to realizing it. Composition of Taiwanese Households

It is not size in itself that distinguishes the Chinese family system from its counterparts in other countries. Nor does it mean a great deal to know about the size of family or household groups—even when we know their distribution from large to small—unless some­ thing can be discovered about their composition. Kinship would not be such a preoccupation in the minds of Chinese if it did not include other things that we have accustomed ourselves to regard as "non-familial." All that the factor of size indicates is the extent to which diverse familial roles can be encompassed within household units. In ascertaining these more diverse activities that were carried on in Taiwanese households, it will be helpful to find out some of the various kinds of members that were present. The census schedules made inquiry into the family relationship of each household member, reckoned with reference to the head of the household. When the person was not a relative of the head, some further statement was still required—whether hired help, domestic servant, or simply a "lodger" in the house. The trouble is that the

THE SETTING FOR FAMILY LIFE

greatest part of this information was discarded by the time the census reached the stage of publication. The work of tabulating and pub­ lishing the wealth of material that was collected on people's family affiliations—which would have entailed tremendous cost—was ap­ parently never carried out. With the 200-odd different positions of kinship that are recognized in the language, and the nine proverbial generations through which they might be counted, probably most of the population of Taiwan would prove to be mutually related if their full genealogies were known. Consequently, the tables that resulted from these censuses indicate relationship only in the in­ definite sense of "family members," without giving any clear idea of what this term actually meant.5 Only two censuses, those of 1920 and 1930, published information on household composition that can stand comparison in detail.® The membership of "ordinary" households at both dates is shown in Table 46, in terms of the average number of persons per household, in each census category. The totals, 5.45 and 5.81 respectively, cor­ respond to those given in Table 45 above. Taiwan's households were made up mostly of "family members," or relatives of the house­ hold heads; the entire group of related persons is represented by both of these two categories combined. Non-related occupants were rare—averaged over all households, they were so few that for most purposes they may be thrown together into one group, "Other." One member of each household was expected to be entered as the "head" in the census enumeration. This designation meant the person who actually represented the household with the outside world and who held authority in making joint decisions that affected the group as a whole. The category was well adapted to the situation, for the same person would as a matter of course assume responsi­ bility to fill out the household's census report. Almost all of the heads were men; but in view of the strong ideal that this should be 5 The directions for filling out the forms in the census of 1920 stated that it should be interpreted to mean "lineal relatives," and we know from the general features of the family system that kinship in this sense would ordinarily be restricted to relatives on the male side only. But no more definite instructions were given on this point by the census regulations. See Census of 1920, Descriptive Report with Appendix, p. 15 of descriptive section. The meaning of "lodger" (or "guest," or "visitor," as the term might also be trans­ lated), is still unclear; apparently it was a catch-all for all people who did not fit into the other categories. β The census of 1915 contained some similar household tables, but these have not been available for this study. Those published for 1920 are the most complete and de­ tailed of all, and will be drawn upon most heavily.

THE SETTING FOR FAMILY LIFE TABLE 46 COMPOSITION OF TAIWANESE HOUSEHOLDS, 1920 AND 1930

Average Number of Persons per Household, by Status in the Household, and Sex

All Household Residents Heads Family Members Occupied a. InHouseholdI b. Outside J No Occupation Employees Domestics "Lodgers"

Change 1920 to 1930 (1930/1920 χ 100) Males Females Both Sexes 1930

1920 Total

Males Females

Total

5.45 0.99 4.36

2.79 0.87 1.85

2.66 0.12 2.51

5.81 0.99 4.71

2.95 0.86 2.01

2.87 0.12 2.71

1.48

0.85

0.63

2.88 0.05 0.01 0.04

1.00 0.05

0.57 0.27 1.17 0.05

«

1.88 e β

0.96 0.35 3.41 0.05

0.02

0.01

0.06

0.03

0.38 Ί C 0.08 2.24 0.01' » 0.03

107 100 108 89 118 110

e Indicates less than 0.005. Due to rounding of figures, sub-totals occasionally do not add to the correct totals. Source: From Census of 1920, Detailed Tables, Table 36 and Census of 1930, Statistical Tables, Total Island, Table 5.

the general rule, there is significance in the fact that more than ten per cent of all households were managed by women. Several signs, which can be discussed somewhat later, suggest that it is not always looked upon as out of the way for women to attain this position. At the moment of enumeration, Taiwanese had more households than heads of households, or not enough heads (regardless of sex) to go around. In this detail Taiwan was different from both Korea and Japan, where there was an inflexible requirement that each household must have one of its members counted in this role. The difference apparently came from a peculiarity of administration, be­ cause for some reason the Japanese did not apply the same regula­ tion in Taiwan. This is not to say, however, that certain households were "headless"—the excess of households over heads meant either that some heads maintained more than one establishment (which might be possible through concubinage), or that the person in this position was in certain instances temporarily absent from home (in which case he would be counted merely as a member of some other

THE SETTING FOR FAMILY LIFE

household, wherever he might be).7 This discrepancy is present in all the tables that follow, but is usually very small and can be ignored. Of the remaining family members in both 1920 and 1930, those who had an occupation were outnumbered by those who had none. There is nothing unusual in this, for the family nearly everywhere is conspicuous for the amount of dependency that it regularly sup­ ports. The generality of this fact is partly due to nothing but the generality of relying on families to bring forth and nurture new generations of adults, which all known societies have done in greater or lesser degree. But dependency is especially prominent in Taiwan and similar countries: children there are very numerous, and they are joined by the aged in partaking of the family's resources on a non-productive basis. The support of each person above as well as below the limits of the working life falls completely on his kins­ men—this is an expectation enforced throughout the society, and it concentrates practically the entire burden on the resources of individual households.8 Excluding the heads of households (who in almost all cases were employed), nearly three-fourths of all other family members had no recognized occupation in 1930—the fraction being less than this for males, somewhat higher for females. Non-producing members took up a large part of family groups in more than just a static sense; people with no occupation also accounted for the major part of en­ largements in household size that accompanied population increase. Even though the period of comparison here is short (the decade be­ tween 1920 and 1930), it is clear that on the average the largest gain per household was in this large category (see the final column in Table 46). In other words, the institutional system translated the growth of population into an increase of dependent family mem­ bers as well as an increase of families, and turned the greater Ioni This second possibility is more plausible, since the excess was greatest where heads were most likely to be absent—in Prefectures with most emigration, outside of agri­ culture (as compared with agricultural households), and in those households with larger numbers of non-relatives. The same was true of Japanese living in Taiwan. β Even though this burden may be mitigated by the fact that old people are not very numerous, and that children are put to work, high mortality usually ensures that a high proportion of infants will survive only long enough to claim their full share of dependency, and will die before contributing any large measure of productive effort. Moreover, countries that are faced with this problem also lack the extra-familial pro­ visions (schools, labor-saving equipment) that many industrial countries have utilized to offset this burden.

THE SETTING FOR FAMILY LIFE

gevity discussed in the previous chapter into additional old and young members of non-working ages.® Family Membership in Households Not all households had the variety in their membership shown by the average of all of them in Table 46. In fact, it was not common for households to contain non-family members at all, and the in­ ternal affairs of most households did not involve persons other than near relatives. Owing to the indefinite criterion of family relation­ ship in the census data, no more than a crude approach can be made to discovering what special or unusual features were present in those families that incorporated "outsiders" into their households. The traditions that shape Chinese familial life are explicit as to why large families are desirable and how they should be achieved. Their foremost concern is that of descent, of preserving the con­ tinuity of a particular family indefinitely into the future. The most emphatic rule of descent is that it should be patrilineal, that a family line may be perpetuated only through its male side. There are no notable exceptions to this rule; in cases where it cannot be followed to the letter, its purpose is observed by means of adoption, changing of surnames, or even reversing some of the usual procedures in mar­ riage, as we shall see in the next chapter. Since the exact relationships are not shown, it is never possible to be certain just which family relatives were included in a household of a given size. But the nuclear group might be described as husband, wife, and unmarried children, for the kinship system does not cus­ tomarily provide these people (or such of them who remain alive and present) any alternative place of residence. With this group alone, much of which was usually made up of dependent children, the actual size of most households was not large enough to leave room for either additional relatives or "outsiders." (See Fig. 30, above.) The goal of still larger families, which was realized by a restricted though substantial part of the population, could not have been at­ tained without encountering special difficulties imposed by the fam­ ily system itself.10 But any further enlargements of a family could be 9 There was some change in the standard of what constituted employment for women between these two censuses, and this accentuated slightly what otherwise would be the pattern of change in Table 46. The same pattern results, perhaps less strongly, if only males are used in the comparison. 10 Of course, the parents of the husband or other relatives of the father's (or higher)

THE SETTING FOR FAMILY LIFE

made in only one way—by retaining in it two or more brothers, with their nuclear family groups, if any. In achieving the extended fam­ ilies that have been cherished in Chinese tradition this is the crucial step, for there are counter-pressures that tend to discourage it. Col­ lateral ties are known to be relatively weak among Chinese—sons do not have bonds of the same strength with each other that they have with their fathers. These ties are especially difficult to main­ tain after the death of a father, for sons may exercise individual claims to shares in the family property. These claims sometimes lead to division of the family and its property even before the father dies; they are intensified as soon as the brothers are married. Extend­ ing the family unit to any considerable size implies some ability to coordinate these conflicting interests and requires leisure for devot­ ing attention to managing its collective affairs, not to mention spe­ cial resources to support its existence on a more elaborate scale. Ob­ viously it was unusual to succeed in preventing family division along such lines as these. If everyone had succeeded, the number of house­ holds in Taiwan would have remained about the same, and the entire increase of population would have gone toward enlarging those that were already there. Although this is not what actually happened, the fact that larger households in Taiwan increased in proportion after 1905 (see Fig. 30) may be a sign that the oppor­ tunities for staving off the division of families were improving. Changes in the size distribution of households are probably an accurate indicator of the rise and fall of family fortunes, if not a very illuminating one. For in reality, when one speaks of ordinary households among Taiwanese, it is practically the same thing as speaking of family groups and nothing else. Nearly everyone who lived in a household of more than one member was a blood relative of the head. In these "co-resident" households, excluding for the moment those Taiwanese who lived alone, over 98 per cent of all Taiwanese in 1920 were related to the household head (if the head generation might be added, for they had claims of support that could not be disre­ garded—one of them might even be the head of the household. But Taiwanese did not live long enough to allow more than three lineal generations to be represented for very long at the same time in the same family. On the other hand, daughters ordinarily ceased to be included in a family after marriage, since the rule of male descent worked in reverse for them—they were ex­ pected to enter the families of their husbands, where in most cases they were found in census enumerations. Though the actual practice in these matters was less uniform than has usually been supposed, it remains true that the expected and preferred ar­ rangement is that outlined here.

THE SETTING FOR FAMILY LIFE

TABLE 47 EXTENT OF FAMILY RELATIONSHIP IN TAIWANESE HOUSEHOLDS, 1920 Persons in All (Ordinary) Households, Those Living Alone, Those Living in Households with Others ("Co-Residents"), Co-Residents Who Were Mutually Related, and the Per Cent Who Were Related. By Type of Household All Household Residents

All Households a. AllTypes 3,454,956 b. Purelyfamily c. Mixed Households with Head Occupied a. Alltypes 3;ill,558 b. Purelyfamily c. Mixed Households with Head Not Occupied a. AllTypes 343,398 b. Purelyfamily c. Mixed Head Occupied in Agriculture* a. AllTypes 2,230,098 b. Purely family c. Mixed Head Occupied Outside Agriculture a. AllTypes 881,460 b. Purelyfamily c. Mixed Head Occupied in Manufacturing a. AllTypes 212,380 b. Purelyfamily c. Mixed Head Occupied in Commerce a. AllTypes 242,048 b. Purely family c. Mixed Head Occupied Elsewhere Outside Agriculture a. AllTypes 427,032 b. Purelyfamily c. Mixed

Persons Living Alone

Mutually Related Fer Cent of Co-Residents Co-Residents Co-Residents (Including Head Mutually (1)-(2) of Household) Related

98.2

3,403,262 3,093,000 310,262

3,343,175 3,093,000 250,175

3,067,141 2,788,813 278,328

3,012,945 2,788,813 224,132

100.0

336,121 304,187 31,934

330,230 304,187 26,043

100.0

14,929

2,215,169 2,033,653 181,516

2,189,415 2,033,653 155,762

98.8 100.0 85.8

29,488

851,972 755,160 96,812

823,530 755,160 68,370

100.0

204,833 173,210 31,623

194,852 173,210 21,642

100.0

6,991

235,057 192,816 42,241

223,124 192,816 30,308

94.9 100.0 71.8

14,950

412,082 389,134 22,948

405,554 389,134 16,420

98.4 100.0 71.6

51,694

44,417

7,277

7,547

* Includes Forestry. Source: From Census of 1920, Detailed Tables, Table 36.

100.0 80.6

98.2 80.5 98.2 81.6

96.7 70.6 95.1 68.4

THE SETTING FOR FAMILY LIFE

is also counted in the figure). This held true in almost the same degree throughout some of the major portions of the society— whether or not the head had a regular occupation, and on the aver­ age regardless of the particular occupation of the head (see Table 47). Closer inspection reveals that the apparent similarity in member­ ship existed chiefly through a preponderance of households made up exclusively of family members, with no outsiders ("purely family households" in Table 47). Of all "co-residents," over nine-tenths lived together in households of this kind, and the case was the same in all the major occupational groups. Whatever differences arose in each occupational group as a whole, therefore, stemmed from a minority of households containing families together with non-relatives—the "mixed households." It is only within this segment that the principal characteristics distinguishing various groups in the population can be found. While four-fifths of all people in even these "mixed" units consisted of family relatives, the proportion was larger in those households dependent on agriculture, and con­ siderably smaller where the head was employed elsewhere (espe­ cially in manufacturing and commerce) (see Table 47). There were different ways of becoming a member of a household, according to the kind of member one was to be, and some of these differences are reflected in the balance between male and female in households of various sizes (see Table 48). Relatives entered a fam­ ily unit through circumstances well covered by institutional rules or procedures—most commonly by birth or by marriage. Neither of these channels permits much latitude of choice to the parties most directly involved—birth is non-voluntary for those being born, and marriage is usually an alliance arranged by persons other than the partners being married. Hence marked sex disparities should not be expected in households that are formed on the basis of kinship alone. However, these "normal" paths of entry were not the only ones, though the predominance of purely family households very nearly made them so. Men obviously had moved out of many family units, especially the smaller ones, and into mixed households of various sizes—not conspicuously into the larger ones. A considerable share of these men must have been single, for departure accompanied by a wife would not have altered the sex balance as between larger and smaller households. Although very few households contained any outsiders, the larger the unit the less likely it was to consist solely of related persons. It

THE SETTING FOR FAMILY LIFE TABLE 48 FAMILY RELATEDNESS AND SEX COMPOSITION OF HOUSEHOLD RESIDENTS

By Size of Household, 1920

Persons per Household

Total Taiwanese Residents (in Ordinary Households)*

51,694 110,618 242,469 385,756 465,810 461,430 394,282 298,984 219,402 162,460

11-15 409,643 16-20 142,703 21-25 59,615 26-30 24,760 31 and above 25,306

1

2 3 4 5 6 7 8 9 10

Males per 100 Femalesf Per Cent in Purely Mixed Households, Ml Family Each Size Households Households

Mixed Households

4.2 3.4 3.8 4.6 5.9 7.6 8.4 10.8 13.1

421 122 100 99 100 103 106 107 107 106

116 98 97 99 102 104 106 105 104

407 186 144 132 123 122 123 121 121

16.4 22.5 26.9 33.2 41.6

106 105 105 106 111

103 102 104 105 104

120 112 109 107 121

—

—

—

° This column was computed by multiplication, and its total is 24 persons less than the totals published in other tables. f These sex ratios were computed from figures for the entire population, Taiwanese and other groups together. Taiwanese, however, preponderate. Source: Census of 1920, Detailed Tables, Tables 33 and 35.

might seem inevitable that households that had added outsiders would be larger than others, but the circumstance is less obvious than at first appears: as we shall see shortly, the households of greater size were not larger by this fact alone, for they also had more family members as well, implying that their familial structure must like­ wise have been distinctive in some respect. Only in households of more usual size, say of less than eight or nine persons, was the purely family type nearly universal. Yet it should be noted that more than half of all Taiwanese in households of more than 25 persons were also living in units made up exclusively of their family relatives— incorporating outsiders was not essential in order to achieve large size.

THE SETTING FOR FAMILY LIFE

Persons Living Alone For a people who take family matters as seriously as they do, Tai­ wanese appeared to forego to a remarkable extent the relationships that grow out of residing together in some kind of household group. The proportion who lived alone does not look especially large—1.5 per cent in 1920 and 1.3 per cent in 1930 (though these people formed a much larger figure when counted as households, as in Fig. 30). But this meant that more than 50 thousand Taiwanese in 1920, and 56 thousand in 1930 were out of immediate contact with a family group at the time of the census (see Table 49). TABLE 49 TAIWANESE PERSONS LIVING ALONE, 1920

("Single-Person Households.") By Occupation Persons Living Alone

Per Cent Total

Per Cent Female

Total Occupied Not Occupied

51,694 44,417 7,277

100.0 85.9 14.1

21.5 13.9 68.1

By Occupation: Agriculture Outside Agriculture a. Manufacturing b. Commerce c. Other

14,929 29,488 7,547 6,991 14,950

28.9 57.0 14.6 13.5 28.9

15.7 13.0 23.1 7.7 10.3

Source: Census of 1920, Detailed Tables, Table 36.

It should not be supposed that for this reason they were out of touch with the kinship system. Fourteen per cent were listed as "not occupied" in 1920, most of whom were probably retired parents without sons or living apart from them. The remainder may have included people temporarily absent from home, perhaps even some of the household heads who were not enumerated with their families. By their livelihood they do not appear to have been a very repre­ sentative group, for most were spread through miscellaneous occu­ pations outside of agriculture, but were rarely found among farm­ ers. Only one-fifth were women, still fewer among those who were self-supporting.

THE SETTING FOR FAMILY LIFE

Without more knowledge about these people there is little that can be done except to remark on their numbers, and perhaps to entertain suspicions that many of them were counted apart from familial households due to arbitrary application of census rules, when they actually were more closely affiliated with some group of relatives. Having the information in Table 49, however, we can separate these relatively isolated individuals from those who were living in constant association with other household members, and thereby gain a better picture of the patterns that arose from this association. This is the meaning that will be given to "co-resident" households in the subsequent discussion. The Composition of Co-Resident Households The average number of persons living together in all "co-resident" households was larger than the figures in Table 46 above.11 The average size of purely family households was almost the same—"out­ siders" were so few that their total number was scarcely noticeable in the figures for households of all types together (see Table 50). But the composition of all households resembled that of purely family households only because nearly all were of the latter type. Mixed households, considered by themselves, were distinguished by more than simply the presence of "outsiders." On the average they also had more members who were family relatives (see the last col­ umn in Table 50). Being larger in both categories brought their av­ erage total size up to 8.1 persons, as compared with 5.7 in purely family households. Outsiders thus were not incorporated into fam­ ilies of the usual sort, but into families that were already exceptional for their large size. In addition, these non-relatives were actually more of a group to be reckoned with than it first appeared; they were distributed so unevenly that they formed a considerable part of the households where they were found.12 The picture of membership for all households was dominated in a similar manner by those households with the head actively em­ ployed. Since most fell into this category, the figures for all house­ holds in Table 46 or Table 50 are little more than a representation of those whose heads had some occupation. The case was the same 11 The 1920 census was the only one that allows this distinction to be made. There­ fore, when comparisons are to be drawn between 1920 and 1930, these "single-person households" must be restored. 12 Some "outsiders" formed households that were nothing but business enterprises, usually small handicraft establishments, with no nucleus of family members. Most, however, were incorporated into households already organized as family units.

THE SETTING FOR FAMILY LIFE TABLE 50 COMPOSITION OF TAIWANESE HOUSEHOLDS-TOTAL, HEAD OCCUPIED AND HEAD NOT OCCUPIED, 1920

Average Persons per Household, by Position in Household, Type of Household, and Sex. (Persons Living Alone Are Excluded)* Households of Purely Family All types Households Mixed Households Both Both Both Sexes Males Females Sexes Males Females Sexes Males Females

1. All Households All Members Head Relatives Others

5.8 1.0 4.7 0.1

3.0 0.9 2.0 0.1

2.9 0.1 2.7 0.0

5.7 1.0 4.7

2.9 0.9 2.0

2.8 0.1 2.7

—

—

—

2. Households with Head Occupied All Members 5.9 Head 1.0 Relatives 4.8 Others 0.1

3.0 0.9 2.0 0.1

2.9 0.1 2.8 0.0

5.8 1.0 4.8

2.9 0.9 2.0

2.8 0.1 2.7

—

—

—

3. Households with Head Not Occupied All Members 5.S Head 1.0 Relatives 4.2 Others 0.1

2.4 0.5 1.8 0.1

2.9 0.4 2.4 0.0

5.1 1.0 4.1

2.3 0.5 1.8

2.8 0.4 2.3

—

—

8.1 1.0 5.6 1.6

4.5 0.9 2.4 1.2

3.7 0.1 3.2 0.4

8.1 1.0 5.6 1.6

4.5 0.9 2.4 1.2

3.6 0.1 3.2 0.4

8.1 1.0 5.6 1.5

3.9 0.5 2.4 1.0

4.3 0.5 3.3 0.5

* Rounding makes some figures fail to add up to the totals that are shown. Source: From Census of 1920, Detailed Tables, Table 36.

for both the purely family and the mixed units. Where the head was not occupied, however, there were certain differences. For one thing, these had more females and fewer males. In purely family households, this shortage of males made the average total size smaller than in those where the head was employed (see Table 51). Mixed house­ holds with the head not occupied, which also had higher proportions of females, were as large as they were elsewhere; but here the de­ ficiency of males for some reason was made up by non-family mem­ bers. For example, almost half of the domestic servants (male and female) found in all Taiwanese households appeared in those whose heads were not occupied—they averaged more than one for every three households.

THE SETTING FOR FAMILY LIFE

TABLE 51 COMPOSITION OF TAIWANESE HOUSEHOLDS, BY OCCUPATION OF HEAD, 1920

Average Persons per Household, According to Position in Household, Taiwanese Households Only. Persons Living Alone Are Excluded. Total, Both Sexes* Households of All Types

Purely Family Households

Mixed Households

6.3 1.0 5.3 0.1

6.1 1.0 5.1 —

9.3 1.0 7.0 1.3

2. Head Occupied Outside Agriculture All Members 5.1 Head 1.0 Relatives 3.9 Others 0.2

4.9 1.0 3.9 —

6.6 1.0 3.7 1.9

3. Head Occupied in Manufacturing All Members 4.9 Head 1.0 Relatives 3.7 Others 0.2

4.7 1.0 3.7 —

6.5 1.0 3.5 2.1

4. Head Occupied in Commerce All Members Head Relatives Others

5.6 1.0 4.3 0.3

5.3 1.0 4.3 —

7.0 1.0 4.0 2.0

5. Head Occupied Elsewhere Outside Agriculture All Members 4.9 Head 1.0 Relatives 3.8 Others 0.1

4.8 1.0 3.8 —

6.0 1.0 3.3 1.7

1. Head Occupied in Agriculture All Members Head Relatives Others

* Rounding makes some figures fail to add up to the totals that are shown. Source: From Census of 1920, Detailed Tables, Table 36.

THE SETTING FOR FAMILY LIFE

Just as in many other instances, the greatest contrast among those with head occupied was that between farming and non-farming households. Those directly dependent on agriculture averaged over six persons each, larger than the rest. The greater part of farming households, however, were purely family establishments (cf. Table 47, above), which did not exceed other familial households in size by any large margin. Many of the largest families, therefore, must have been concentrated in that portion of farming households in which outsiders were mixed with family members, which reached an average size of 9.3 persons each. Agriculture thus provided the most favorable setting for realizing the goal of extended and joint fam­ ilies, although one suspects that large farming households may have been partly due to lack of employment opportunities elsewhere. But though large families and the presence of non-family mem­ bers tended to go together in Taiwanese agriculture, the opposite was true in most other parts of the society. Mixed households out­ side of agriculture were actually smaller as family groups than their counterparts consisting of family members only. At the same time, mixed households outside agriculture had substantially more out­ siders than mixed farming households. While the mixed households in agriculture were large by virtue of the cohesion provided by fam­ ily ties, they were larger outside of agriculture in spite of it. Whatever weakening of one's own familial connections was asso­ ciated with working or living in the household of someone else, it was not equally available to persons of both sexes. The Chinese family system makes little provision for women to live away from the family of either their parents or their husbands, and in Taiwan hardly any of them did so.13 There were few women anywhere in Taiwan living in households where they were unrelated to the head— less than one per cent of all those counted in 1920 (see Table 52). The fraction that lived in quasi-households was negligible. Though men were also limited in their opportunities for individual achieve­ ment unencumbered by family obligations, such restrictions were nearly unavoidable in the case of women. Even in those households shown in Table 50 and in Table 51 that did contain some outsiders —i.e., the mixed households—less than one-fifth of the women were 13 Many of these women must have been merely the wives of men who were staying in the households of others. Most of the non-related women in mixed households were in the ambiguous category of "lodgers," which leaves us without any clear indication of the circumstances of their position.

THE SETTING FOR FAMILY LIFE TABLE 52 "OUTSIDERS" IN TAIWANESE HOUSEHOLDS, 1920

Per Cent of All Members Who Were Not Relatives of the Head, by Sex, and by Occupation of Head. (Heads of Households Are Included, Persons Living Alone Excluded)

A. Households of All Types All Households Households with Head Occupied Households with Head Not Occupied Head Occupied in Agriculture Head Occupied Outside Agriculture Manufacturing Commerce Other Mixed Households All Households Households with Head Occupied Households with Head Not Occupied Head Occupied in Agriculture Head Occupied Outside Agriculture Manufacturing Commerce Other

Males

Females

2.6 2.7 2.5

0.9 0.8 1.1

1.8 4.8 7.4 7.3 2.1

0.5 1.7 2.2 2.7 1.1

26.8 27.0 25.2

10.3 10.0 12.3

21.0 37.1 40.5 36.0 34.2

6.2 18.1 17.4 17.3 20.4

Source: From Census of 1920, Detailed Tables, Table 36.

not family members, whereas outsiders usually were at least twice as frequent among the men. Participation by Women in Household Roles

The Chinese are notable in not allotting to women a status of any particular prominence or responsibility in the society at large. As a result, their participation in non-familial activities is very re­ stricted, and the evidence of previous chapters has not often shown them to be engaged in affairs outside of their family groups. Since they are not found anywhere else, it is of some interest to discover what is possible about the position of women in the household. Significantly, in the households of various types shown here, women were distributed almost everywhere in about the same pro-

THE SETTING FOR FAMILY LIFE TABLE 53 PROPORTIONS OF FEMALES IN TAIWANESE HOUSEHOLDS, 1920

By Type of Household and Occupation of Head Households Purely Family Mixed of All Types Households Households

A. Per Cent Female: Family Members* All Households Households with Head Occupied Households with Head Not Occupied Head Occupied in Agriculturef Head Occupied Outside Agriculture a. Manufacturing b. Commerce c. Other Per Cent Female: Head of Household All Households Households with Head Occupied Households with Head Not Occupied Head Occupied in Agriculturef Head Occupied Outside Agriculture a. Manufacturing b. Commerce c. Other

49.6 49.0 54.7

49.6 49.0 54.6

50.1 49.3 56.4

49.0 49.1 50.0 49.4 48.5

48.9 49.3 50.4 49.6 48.6

50.2 47.3 46.9 47.9 46.6

11.2 7.1 45.1

11.3 7.2 45.1

10.6 6.5 45.7

7.4 6.4 12.8 3.8 4.6

7.3 6.5 13.5 3.7 4.6

7.3 5.5 7.5 4.3 4.8

° The head of the household is included as a family member, f Includes Forestry. Source: From Census of 1920, Detailed Tables, Table 36.

portion to men (see Table 53). Slightly less than half of the member­ ship of most households was female, and it made almost no difference what was the family's source of support.14 Despite the association between sex composition and household size found in Table 48, the degree of subordination of women apparently did not vary as be­ tween any of the principal lines of employment—women did not gain their positions in Taiwanese households through independent achievement, but chiefly through their attachment to men. ι* The proportions of all household residents who were women is not shown in Table 55. Since women were nearly all members of the family, their percentages among family members is practically the same thing, except in mixed households.

THE SETTING FOR FAMILY LIFE

This uniformity was due in large measure to the fact that almost all women were members of their households by virtue of kinship, and hardly any of them stayed as non-relatives in the families of others. True, slightly over 20 per cent of all persons who were counted as outsiders were women; but this still did not amount to a large number, and it has already been pointed out that many must have been wives or children of men in this status. Though on the average they were found in nearly equal numbers with men, there is no chance that women occupied the same posi­ tions in their families. Chinese ideas of propriety in behavior are never more consistent than in creating a distinction in familial roles based on a person's sex. The patriarchal ideal of authority in Tai­ wanese families is illustrated by the infrequency with which women became heads of households. Among those of the household heads who were employed (that is, practically all of them), men exceeded women by nine to one; in all but seven per cent of these households it was reported that the affairs were managed by men. In the minor­ ity of families where the head was not employed, this responsibility fell almost as often on women as on men—45 per cent of these house­ holds were managed by women. The only way for most women to accede to this position was through the death of their husbands. Therefore the concentration of female heads of households among the households with no apparent source of family support most likely represents those widows who were not taken in by their hus­ bands' relatives. Family Membership and Employment Questions of livelihood, which are important to all peoples, are treated with a particular seriousness by Chinese. They are usually regarded as questions of household management, and thereby be­ come closely interlocked with the ideas and values of the family system itself. When this happens, a large part of a society's economic life is likely to be sealed up in self-sufficient family units, and to lack the aspect of inter-dependence that characterizes the activities of individuals in the industrial regions of the world. The households of Taiwan were the scene of most of the effort that went into providing support for the population, and hence are the most appropriate place to look for traces of its organization. A good indication of the contribution to household support in differ­ ent classes of households is given by the extent of regular work by

THE SETTING FOR FAMILY LIFE

male residents. The work of women may have been equally impor­ tant; but the indices for assessing it have the same shortcomings in making such comparisons that they had in measurement of the labor force—female labor was almost entirely contained within household units, and the census statistics do not represent it accurately at all, but only the willingness to acknowledge and report it as such. Work by men was looked upon more readily as maintaining the household in relation to the external society, despite a similar tend­ ency for it to be in actuality an internal matter. Male family mem­ bers seemed to have regular employment in almost the same propor­ tion regardless of the occupation of the head of the household. Slightly over 60 per cent of them reported some occupation in 1920 if the head is counted. This, of course, is only what is implied by the findings in earlier chapters, for the figure is close to that fraction of the entire male population in working ages. Although the av­ erage size of family varied among different groups, the share of family members' services taken by full-time work did not. The one exception to this rule large enough to appear in Table 54 occurred in families engaged in commerce, which for some reason held more non-employed men than families in any other non-agricultural type of work.16 But if leisure was supported to any appreciable extent in Taiwanese families, there is little sign that it was concentrated in any particular field of livelihood. It will make the comparison more complete if we also adopt an alternative view of dependency. Since in households with head oc­ cupied the head by definition cannot be a dependent, it is only fair to compare family members in commercial and other households after the family heads are excluded. The remaining males are the ones whose conduct is subordinate to the authority of the head, and who represent the part of the population most likely to have leisure status where a family's wealth will permit. It would appear that farming families, which had more of these subordinate members to support, put their services to much greater use. Only 35.5 per cent of subordinate males in commercial households reported employ­ ment in 1920 (35.9 per cent in purely family households). Other families tended to be distributed in between (see the columns "Ex­ cluding Head" in Section B, Table 54). 15 Whatever the explanation for this peculiarity, it stemmed from an unusually large number of non-occupied men rather than abnormally few who had an occupation, since commercial families were uncommonly large in every respect as compared with other non-agricultural households.

THE SETTING FOR FAMILY LIFE

Families with unoccupied heads, which in several respects formed a class by themselves, were not so distinct when it came to employ­ ing their subordinate male family members. When the head is in­ cluded in the computation, the fact that he (or she) was not occu­ pied gives an appearance of dependency far greater than that found anywhere else in the population. Excluding the head (in Table 54) reveals that more of the remaining family members were employed than in any other field of livelihood except agriculture. Unfortu­ nately there is no way of knowing what sorts of work they did; if TABLE 54 EMPLOYMENT AND DEPENDENCY IN TAIWANESE HOUSEHOLDS, 1920

Male Family Members Who Reported a Regular Occupation, and Those Who Reported None, by Type of Household A. Average Male Family Members Per Household, Occupied and Dependentf

Households of All Types

All Households a. Occupied 1.74* b. Dependent 1.15 Households with Head Occupied a. Occupied 1.85 b. Dependent 1.11 Households with Head Not Occupied a. Occupied 0.86 b. Dependent 1.49 Head Occupied in Agriculture$ a. Occupied 2.01 b. Dependent 1.18 Head Occupied Outside Agriculture a. Occupied 1.52 b. Dependent 0.97 Head Occupied in Manufacturing a. Occupied 1.46 b. Dependent 0.89 Head Occupied in Commerce a. Occupied 1.56 b. Dependent 1.11 Head Occupied in Other NonAgricultural Occupation a. Occupied 1.54 b. Dependent 0.93

Purely Family Households

Mixed Households

1.74* 1.13

1.84' 1.43

1.84 1.09

1.96 1.36

0.86 1.46

0.85 2.04

1.99 1.15

2.32 1.66

1.53 0.97

1.48 0.97

1.45 0.89

1.47 0.90

1.57 1.11

1.49 1.13

1.54 0.94

1.47 0.81

THE SETTING FOR FAMILY LIFE TABLE 54 (continued)

. Per Cent of Male Family Members Occupied

Households of All Types Ex­ In­ cluding cluding Head Head

Purely Family Households Ex­ In­ cluding cluding Head Head

Mixed Households Ex­ In­ cluding cluding Head Head

60.2

45.8

60.6

46.1

56.3

42.4

62.5

45.6

62.8

45.8

59.0

43.1

36.6

47.7

37.2

48.7

29.5

36.2

Head Occupied in Agriculture 63.0 Head Occupied Outside Agr1. 61.2 Manufacturing 62.0 Commerce 58.4 Other 62.3

48.0 38.4 40.0 35.5 39.3

63.4 61.3 62.0 58.7 62.2

48.2 38.6 40.2 35.9 39.3

58.3 60.4 61.9 57.0 64.5

45.7 36.4 38.4 33.0 40.3

All Households Households with Head Occupied Households with Head Not Occupied

* Household heads with no occupation are of course not included. All figures of average persons per household have been rounded, and sometimes will not add to the totals that are shown in other tables; they also will not always produce the exact percentages shown in the lower section of this table. f "Dependent" refers to persons reporting no occupation in the census. J "Agriculture" includes Forestry. Source: From Census of 1920, Detailed Tables, Table 36.

many of the unoccupied heads were widows who had not re-married, it may be surmised that these households probably had representa­ tives in most occupational groups. But the most pervasive distinction in engaging in work or leisure within families appeared in those that incorporated outsiders to as­ sist in the work of the household—the mixed households—which made less use of labor by family members than households consisting of family members only. In all but one of the occupational classes of Table 54, employment was less common among members of the families living in mixed households, whether the heads are counted or not. Households with heads not occupied were no exception. Even their subordinate male family members were put to work less often— when not dependent solely on their own resources of manpower, such families supported the greatest amount of leisure that can be found in any of these groups. Differences of Occupational Status

The variation that these data show in the composition of family groups was recorded at a date before any strong force of Japanese

THE SETTING FOR FAMILY LIFE

government policy could have been felt in this sphere, and so it must have been a pattern of long standing among the Taiwanese. Though there is no way of knowing how much the variation amounted to between individual families, between various sources of livelihood it was of a very restricted nature. One of the principal features of the Chinese social order, however, is that there has been very little variation in the general notions that people in different stations hold about what forms of conduct are desirable—the standards of behavior within one's family, or with members of related families, have been remarkably similar from group to group, whether or not these ideals are completely fulfilled in fact. It has also been charac­ teristic of this system through many historic periods that it has not given rise to rigid status and class distinctions that became heredi­ tary and remained fixed over time. The commonest sources of pres­ tige have been associated with symbols that leave some room for personal achievements, and not simply privileged birth, in their at­ tainment—the adoption of genteel as apart from vulgar pursuits, a highly esteemed and arduous classical education, and observance of traditional forms of ritual that are inappropriate to commonplace affairs. Wealth, in almost any form of livelihood, has conferred a capacity (e.g., through leisure) to practice these ideals, in many ways that poverty denied to the greatest part of the society. Status distinctions of this sort are lines that cut across most if not all of the occupational divisions recognized in Japanese census prac­ tice, though they obviously were present in some more than others (commerce or government service, for example, more than agricul­ ture or fishing). There are further resources, however, to use in find­ ing out the extent of variation within occupations. In the one master table published with the census of 1920, households were also classi­ fied by three additional categories of "occupational status"—"Own­ ers" or "Independent Workers," "Officials," and "Laborers."16 Each household, if the head was occupied, was assigned to one of these groups as well as to a regular occupation, so as to specify more fully the nature of the head's position. The classification is limited in value, for it is not certain that the terms signify anything that is ie The English terms are only approximations of the meaning in Japanese. The sense of "Independents" (Yeh-chu in the Chinese reading of the term) really is closer to the notion of "master" of an enterprise, somewhat like the term "self-employed" in Western census practice. "Officials" (Yi-yilan in Chinese, where it means a lowly and subordinate position) has an honorific connotation in Japanese. "Laborer" (read Laowu in Chinese) is relatively unambiguous, though it is not clear what line divides it from the others.

THE SETTING FOR FAMILY LIFE

consistent from one occupation to another. In agriculture, a person apparently was "independent" if he actually operated a farm (with­ out distinction between landlords and others), and "laborer" if simply working for another. Agricultural occupations had hardly any positions that could be designated as "officials"; practically all of these were found as government officials or professionals of some kind. These terms are so vague that it is useless to be distressed over what they do not reveal. The service they do perform is to separate, within each broad occupational class, the more from the less suc­ cessful families ("independent" as opposed to "laborers," based on the fortunes of the household head), and to distinguish a few of the cases that do not easily fall in either of these (the "officials," wher­ ever they occurred). The households of "laborers," in each of the main classes of occu­ pations, were on the average smaller than those in the same occu­ pations that were more prosperous, and smaller even than those of "officials" (see Table 55). The more prosperous ("independent") households were consistently larger than either of the others, whether they were composed solely of family members or included outsiders.17 Mixed households of "independent" heads were the largest of all, and had smaller proportions of outsiders than mixed households of the two other "status" groups. In other words, the wealthier mixed households exceeded the lesser ones in the same manner that had marked the difference between mixed households and those com­ posed of family relatives only—chiefly through the extension of their families rather than the incorporation of outsiders. N ot merely in the families of farmers, but within any occupational category, the households of the well-to-do were larger than those of the poorer portions of the population, and in this same pattern. Households whose heads were "independently" employed had the smallest proportions of their male family members engaged in fulltime work. But while leisure, as measured by this standard, was especially common in "independent" households whether outsiders were present or not, it was greater in mixed households of this group 17 Here, as before, persons living alone ("single-person households") have been ex­ cluded, and Table 55 represents co-resident households only. Therefore the actual num­ ber of persons per household is less important (e.g., for comparison with later dates in Taiwan or with other data), than the pattern of variation from group to group in Table 55. It happens in this case that most of the single-person households were of the "laborer" group, and so would make the contrast between the two more pronounced if included, by giving the "laborers" a still smaller figure for average household size.

THE SETTING FOR FAMILY LIFE TABLE 55 AVERAGE SIZE OF HOUSEHOLD AND EXTENT OF MALE EMPLOYMENT IN FAMILIES CLASSIFIED BY "OCCUPATIONAL STATUS," 1920

By Occupation of Head and Type of Household

Number of Households

A. Average Persons per House­ hold, both Sexes, by Occupation of Head. All Occupations a. Independents* b. Laborers c. Officials Agriculture a. Independents b. Laborers Outside Agriculture a. Independents b. Laborers B. Per Cent of Male Family Members Who Reported an Occupation All Occupations a. Independents b. Laborers c. Officials Agriculture a. Independents b. Laborers Outside Agriculture a. Independents b. Laborers

Purely House­ Family holds of House­ holds All Types

Mixed House­ holds

388,635 128,768 1,462

6.4 4.4 4.9

6.2 4.4 4.6

8.7 5.4 5.8

309,158 41,314

6.6 4.2

6.4 4.2

9.5 5.5

79,477 87,454

5.7 4.5

5.5 4.5

7.1 5.3

62.0 65.1 54.0

62.3 65.0 52.5

57.9 69.2 60.9

62.6 67.6

63.0 67.6

58.1 65.2

58.8 64.0

59.0 63.7

57.3 70.0

For an explanation of the meaning of these terms, see text. Source: From Census of 1920, Detailed Tables, Table 36. β

than anywhere else (see Table 55, Section B). Apparently these more prosperous households also took in outsiders for different pur­ poses. Though not shown in Table 55, the non-relatives in "labor­ ers' " households were chiefly "lodgers" (whatever might be included under this term); the ones in "independent" households were mostly

THE SETTING FOR FAMILY LIFE

listed as employees, who allowed family members to be released from regular work. Changes Between 1920 and 1930

The decade between 1920 and 1930 produced little change in household or family structure as far as size or other observable fea­ tures were concerned (see Table 46, above). Families on the average became slightly larger than before, but chiefly by additions to their dependent children and possibly older relatives. When the data for 1920 are put back into terms appropriate for comparison with 1930 (that is, by replacing the single-person households in the computa­ tion of averages), it appears that the growth of households was dis­ tributed almost uniformly in different portions of the population— it is necessary merely to add the figure of about 4/10 to the average size of households in each major occupational group in 1920 to ap­ proximate the size found in that group in 1930. The tables showing household composition were much abridged in 1930, and permit only restricted comparison. In many of the features already used to describe their structure in 1920, however, households in 1930 were practically the same. They were made up almost entirely of family relatives; and presumably the contrasts between mixed and purely family households were unaltered. The same uniformity in sex com­ position prevailed among the various classes of occupations, with the same exception in the case of households with non-occupied heads. Since these data contribute nothing to the discussion, they have not been reproduced here. There is more to be gained by comparing economic activities of family members. The burden of dependency was on the average slightly increased between 1920 and 1930, as we have seen (Table 46, above). By 1930, households representing all the main sources of support (i.e., occupations of heads) had smaller proportions of male family members occupied than 10 years before.18 Agricultural households seemed to share less in this rise of family dependency than others (see Table 56). In part the divergence may be because peasant agriculture is better able to provide a secure livelihood for many who could not find steady employment else­ where—especially young boys and older men. We have found, in is How the added burden was shared between households of only family and mixed households cannot be ascertained, though the similarity of other patterns leads one to presume that this one remained nearly the same as in 1920. This was one of the items of detail omitted in publication of the 1930 census.

THE SETTING FOR FAMILY LIFE TABLE 56 CHANGES IN THE EXTENT OF MALE EMPLOYMENT IN TAIWANESE FAMILIES 1920 TO 1930

(Persons Living Alone Are Included) Per Cent of Male Family Members Occupied, Including Head 1920 1930

All Households Households with Head Occupied Households with Head Not Occupied Head Occupied in Agriculture0 Head Occupied Outside Agriculture*

61.0 63.4 36.6

56.9 59.6 36.6

63.4 63.4

60.0 58.5

s The remaining occupational groups are not shown for comparison here, because the tables published in 1930 followed a different scheme of classification for house­ holds—according to the "industry" of the heads. The category "Agriculture" has been made comparable in 1930 by taking those engaged in "Cultivation" only. (A special table published in 1930 shows that almost everyone in this category was also classified as having an agricultural "occupation"; workers in "Forestry," of course, are not in­ cluded.) The group with head "Not Occupied" is the same in both occupational and industrial classifications. Source: From Census of 1930, Statistical Tables, Total Island, Table 5. Data for 1920 from preceding tables.

Chapter IV, that agriculture in Taiwan exerted its greatest attrac­ tion compared with other occupations for men of these ages. Thus men of these borderline ages probably were reported as employed in agriculture, and as dependents elsewhere, and some of this di­ vergence may be simply a difference of standards rather than of work that was performed. But farming families had more use for the services of these people (and of those with physical disabilities); it is possible that the greater readiness to recognize these services—of women as well as of men—reflects their actual importance in con­ tributing to the support of their households. This apparent trend toward creating a gap between agriculture and other parts of the society was not a very impressive one. It is worthy of attention chiefly because it stands alone; there are no other new patterns of dependency to be found between 1920 and 1930, since the other occupational categories cannot be matched at both dates. And, more important, insofar as there was a leisured group or class among Taiwanese, it was still not to any marked ex­ tent occupationally aligned.

THE SETTING FOR FAMILY LIFE

Work Within the Household People's working activities play a part in familial life that is larger than merely the support that they provide. Since they are ordinarily not carried on in isolation, these activities supply one of the con­ nections between the family and the remainder of the society. In industrial countries they sometimes appear to dominate this connec­ tion, and have become one of the criteria that fix a person's station in society. In agrarian countries it tends to be the other way around; the working roles of most adults are carried on as adjuncts to their other familial roles, and depend on kinship obligations for coordina­ tion. This, for example, is almost completely the case in household enterprises, and holds true in varying degrees for such wider market transactions that exist. In Taiwan, the productive effort of most people was in fact con­ fined to work pursued in cooperation with other family members and organized as household duties. This has already been demon­ strated in a negative way as far as women were concerned—the oc­ cupational statistics, in failing to show very much about the work of women, do confirm that they were scarcely ever employed outside of their households. In household enterprises—and practically all farming was conducted on this basis—the head of the household was almost automatically the chief participant. However, since the sta­ tistics give no positive answer in the case of women or household heads, it is in the activities of subordinate male family members who were employed that we have the best indication of the concentration of work within households, and of its variation in different sections of the society. Except for the minority counted as "employees," these remaining men were the ones with most opportunity to find work unrestricted (or unaided) by their kinship affiliations—i.e., in something that might approach a "labor market." Yet as late as 1930, over two-thirds of the Taiwanese men in this group still pursued their occupations in the households of their own families (see Table 57). In reality their labor was even more family-centered than this. In the large majority of households the head had some occupation, and there 76.1 per cent of all occupied male family members worked in the same enterprise. The remaining household heads had no occupation them­ selves, and hence did not operate family enterprises that could claim the services of men in their families. Most of the men in these fam-

THE SETTING FOR FAMILY LIFE TABLE 57 CONCENTRATION OF EMPLOYMENT WITHIN HOUSEHOLDS, 1930

Per Cent of "Subordinate" Male Family Members (i.e., Excluding the Household Head) Who Were Employed Within Their Households (Taiwanese Households Only, Including Persons Living Alone) Per Cent of Employed "Subordinate" Male Family Members Whose Employment Was Inside Their Household

All Households Households with Head Occupied Households with Head Not Occupied Head Occupied in Agriculture* Head Occupied Outside Agriculture Manufacturing Commerce Other

68.2 76.1 21.6 87.3 34.2 33.1 42.2 26.3

* See note to preceding table. Source: From Census of 1930, Statistical Tables, Total Island, Table 5.

ilies who were occupied found employment outside—only 21.6 per cent of them carried on their work in the family.19 Among those households with heads occupied, there were differ­ ences just as great between the fields of livelihood that provided support for the households. Farming families kept 87.3 per cent of their subordinate working males occupied at home. It was their nu­ merical weight in the population that made the figure for all occu­ pations so high—in non-agricultural households it was only onethird (see Table 57). The influence of the more modern parts of the economy is indicated by the fact that families engaged in these fields had the greatest proportion of working-out among their eligible men.20 Of the non-agricultural categories of occupation shown in Table 57, the commercial households came the closest to conforming to the norm exemplified in agriculture. is Literally, the data of Table 57 refer to those working in the "same occupation or same enterprise" as the head of the household. The meaning is not so broad as the literal translation might suggest, because "same occupation" did not mean merely "also in agriculture" or "also in commerce" or some other large category. It means the same type of work in considerable detail—"also a butcher" or "also a bean-curd maker," for example. In nearly every case it probably meant work in the same family enterprise. 20 Of course, many who worked out may have gone to work for relatives, and still did not enter the labor market.

THE SETTING FOR FAMILY LIFE

Of interest for the sake of comparison are the corresponding data illustrating the variation between different parts of the island (Table 58). In most of the individual Prefectures there were only minor differences in the extent to which employed men in all families (ex­ cluding the heads) performed their work at home. The most popu­ lous sections seemed uniform in the proportion that were so enTABLE 58 CONCENTRATION OF EMPLOYMENT WITHIN HOUSEHOLDS: VARIATION BETWEEN PREFECTURES OF TAIWAN, 1930

Per Cent of Employed "Subordinate" Male Family Members Whose Employment Was Inside Their Households

Name of Prefecture

Taipei Hsinchu Taiehung Tainan Kaohsiung Taitung Hualien P'enghu Total Island

All Households

Head Head Head Occupied Households Outside Reporting No With Head Occupied in Occupied Occupation Agriculture* Agriculture

50.4 72.3 71.0 71.6 72.2 95.1 82.5 64.5

59.6 79.5 79.1 79.5 78.4 95.8 85.0 64.5

81.8 88.5 88.3 88.6 85.7 97.6 91.8 75.1

21.4 34.3 36.5 41.6 45.9 52.0 30.5 54.8

10.5 19.7 25.6 25.7 26.5 69.9 34.1 30.7

68.2

76.1

87.3

34.2

21.6

9

See footnote for Table 56. Source: From Census of 1930, Statistical Tables, by Prefecture (respective Prefectural volumes), Table 5.

gaged; the deviations in the two Prefectures on the East Coast (where hardly anyone worked outside of his household), and in P'enghu Islands, did not affect many people. Taipei Prefecture was the sig­ nificant exception to the rule, for it was the only place with consid­ erable numbers of male workers occupied in extra-familial pursuits. The more detailed columns of Table 58 help to clarify the source of the variation. As the site of the capital city and administrative center of the island, and the portion in which Japanese concentrated their efforts in developing modern conveniences and comforts, Tai­ pei had even before 1930 gained many features of distinction over the rest of the island. It was at this time the transportation hub (built around shipping from the port of Chilung, which later de-

THE SETTING FOR FAMILY LIFE

clined), the home of one-third of all Japanese, the location of the better educational facilities, the commercial area of most importance for trade in other than farm products, and of course the center of most large government and business offices. Taipei Prefecture, there­ fore, had proceeded farthest in the spread of non-agricultural activi­ ties in its population. Even though a substantial group in this popu­ lation does not appear in these tables because they were Japanese, the effect of their presence is evident in the way that non-farming occupations were conducted by Taiwanese—all but 21 per cent of the men in such work (excepting the heads of households and people staying in the households of others) were employed away from their families, a proportion less than any other Prefecture and only half the figure in some. Agriculture was much less affected by the unusual character of other pursuits, but it did not go untouched. Though most farmers in Taipei did their work in their own households, they fell somewhat below the mark set by the rest of the agricultural population.21 Significantly, in all of the other Prefectures (except P'enghu, which was not of much account in size), agricultural households were nearly uniform in the degree to which they retained the use of services by subordinate men in their families—close to 90 per cent of the men in this category who lived with their families.22 The principal differ­ ences from one Prefecture to another when all households are con­ sidered at once, therefore, must have arisen from the prevalence and the mode of organization in activities outside of agriculture, for they practically disappear when farming households alone are compared. There can be little doubt that the family system, through its weight in the conduct of household affairs, has furnished the most impor­ tant social setting for behavior of the Taiwanese. We have not only the testimony of Chinese tradition to support this conclusion, but the corroboration of statistical evidence as well. Practically everyone was a member of some familial household—most were related by Taipei was the site of tea plantations, which, through a greater use of wage labor, may have been responsible for the deviation from the pattern elsewhere in the island. A plantation might employ men (and actually did employ many women) on a more or less regular basis from several farming households, and leave their services free at some seasons of the year for work in their own fields. 22 Naturally this does not include those who had left home to enter some other occu­ pation; these would appear as outsiders in some other households or as persons living alone, perhaps in some cases as separate families.

THE SETTING FOR FAMILY LIFE

kinship, and those who were not, the "outsiders," generally owed their membership to the work that they performed in the families of others. In addition, many strong obligations were observed among people who were related but not living together in the same house­ hold. So in reality the significance of the family system was wider than can be adequately conveyed by the facts of this chapter; it extended beyond the bounds of kinship, in its application to nonrelated household residents, and beyond the limits of individual households, in governing the intercourse between related families. As for those Taiwanese who lived in their own familial households, the best-documented of all their activities is the work that they per­ formed. The extent of male employment in these family groups showed no large differences as between the major classes of occupa­ tions. But leisure, as judged by the abstention from regular work, was apparently more widespread in families of greater affluence, re­ gardless of their particular source of support. And their households were also the ones that made the most use of services of non-family members, which was one of the ways of making leisure possible. It is significant that the work of most men was carried on in house­ hold enterprises inside their own families. In a sense this was a ful­ fillment of traditional exhortations that families should be self-suffi­ cient, but a great deal more was involved at the same time. A society which is family-centered in most respects cannot afford to make very far-reaching exceptions in the case of economic activities: since em­ ployment, if it occurs outside of the household, is a possible avenue for a person to advance in social position independently of family influence, it also is potentially a competing source of loyalty. In this and various other ways, the stability of the social order was buttressed by the force of its own organization as well as through the support of its moral precepts, by making it difficult for people to by-pass the family system in pursuit of their self-interest. Though the span of observation is short and confined to a rela­ tively early period in Taiwan, the facts betray no remarkable sign of yielding in this household structure. And changes that did not alter this fundamental pattern could not amount to very much. The fail­ ure of "colonial development" to be a clearcut departure from the past among Taiwanese should be attributed very largely to the sta­ bility of their familial institutions.

CHAPTER VIII

Patterns of Marriage and Divorce Up το this point, it has been possible to describe the manifestations of the Chinese family system in Taiwan only through an abbreviated view of the families existing at a particular moment—the view which is afforded by enumeration of households in the census. This ap­ proach suffers from several shortcomings. The delineation of what should comprise a family, inherently arbitrary, was one that took account of only those family members who resided together, and thus understated the influence of kinship in the society. Also, the data give but a sketchy indication of family members' character­ istics from which we might deduce more satisfactorily their familial roles (items like age, marital status, and their degrees of relation­ ship to each other). In short, our knowledge of what occurs in Tai­ wanese families is not so well documented as the behavior of their members outside of the family units, and comes largely by inference from this second kind of information. Part of this gap may be filled in two ways from other sorts of data. The conclusions about family composition will be supplemented in this chapter by a discussion of marriage, on which depended the formation of new family units and the continuation of old ones. The next chapter will be devoted to measuring the fertility of Taiwanese, through which their families replenished the population with new members. Like people in other countries, most Taiwanese marry sooner or later. The differences between Taiwan and many Western countries are superficially those of degree—for most Taiwanese men and women marriage comes early and is nearly universal. It ranks as One of the most significant events that can happen in a family, and is at­ tended by attitudes and customs that reflect the importance ascribed to it. The wedding itself is ordinarily solemnized by the most elab­ orate and festive ceremony that a family can afford.1 The idea of marriage is so endowed with importance and virtuous purpose that for a person to remain celibate is usually a mark of unexampled misi There are many descriptions of the conduct of Chinese on these occasions and the ways in which this fits in with other familial life. Cf. Lang, O. (1946), and Levy, M. J- Jr- (1949)·

PATTERNS OF MARRIAGE AND DIVORCE

fortune. In contrast to the West, Chinese culture entertains no al­ ternatives that offer serious competition to the "normal" family roles. TABLE 59 PROPORTIONS OF TAIWANESE EVER-MARRIED, BY AGE AND SEX, 1905 TO 1935

Males and Females Age

10-14 15-19 20-24 25-29 30-34 35 and over 10-14 15-19 20-24 25-29 30-34 35 and over

1905

0.1 10.2 45.7 77.4 89.4 95.3 2.3 47.3 91.6 98.2 99.2 99.7

1915

1920

1925

1930

19S5

0.0 5.7 40.1 72.9 88.1 95.4

Males 0.2 6.3 42.8 76.5 88.6 95.5

0.1 5.3 44.7 78.2 89.7 95.8

0.1 7.3 51.6 82.9 91.3 96.2

0.1 5.7 49.8 83.1 91.9 96.0

0.7 34.7 87.4 96.6 98.5 99.5

Females 0.6 32.8 86.6 96.9 98.5 99.4

0.4 29.4 84.4 96.2 98.3 99.3

0.5 32.6 86.3 96.1 98.0 99.3

0.4 28.1 83.0 95.9 97.7 99.0

Source: Tables of the Taiwanese population by age, sex and marital status in the reports of each census. The "Informally Married," who in 1920 were shown as a subtotal of "Unmarried," have been reinstated here as "Married," in conformity with the census procedure at other dates.

There has been no indication from related spheres of activity already discussed that these pressures to marry were anything but constant and unyielding during Japanese rule in Taiwan. The per­ centages of Taiwanese over age 30 who eventually married remained about the same—all but one or two per cent of women, and all but nine or ten per cent of men. Yet the proportions of Taiwanese who had ever been married did not remain constant at all ages. The younger people of both sexes displayed some signs of altering the time when their marriage took place (see Table 59). So far as this evidence is germane, however, it offers conflicting testimony: the pressures would appear to have been weakened for women under age 30 and to have grown stronger for men. Faced with this apparent inconsistency, we naturally cannot vouch for the inner motives of people about to marry. Nor can we adopt the usual

PATTERNS OF MARRIAGE AND DIVORCE

course in demography and merely say that there was a trend for people in Taiwan to "marry earlier" or "marry later," for this is to say nothing about why men and women apparently were of differ­ ent minds in the matter. But thanks to the assiduity of the Japanese in maintaining Taiwan's population statistics, the problem is open to closer examination with the help of additional data. TABLE 60 SEX BALANCE OF TAIWANESE IN MARRIAGEABLE AGES

Males per 100 Females Enumerated in Ages 15-49 in Successive Censuses, 1905 to 1940 Age

10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 Total, 15-49

1905

1915

1920

1925

1930

1935

1940

122 126 123 122 119 116 112 100 119

113 116 121 123 119 114 107 100 115

109 112 115 120 120 114 108 99 112

107 109 112 114 118 118 110 101 111

105 106 108 111 114 116 114 105 109

103 104 106 108 110 112 114 109 107

104 100 102 104 107 109 110 108 104

Source: Tables of the Taiwanese population by age, sex and marital status in the reports of each census.

It is worth while first to consider the conditions of the marriage market in Taiwan, in terms of the relative supply of men and women available to marry. Since, practically speaking, the Taiwanese were a "closed population," this is one instance where demographic trends in births and deaths may be observed very simply in their bearing on the more complex aspects of behavior. We have already seen that the chances of surviving through infancy and childhood had been extremely poor for girls born before 1895, and that this handicap disappeared only gradually in later years. As a result, by the time they were nearing age 20—the range when most of them married— there were not enough women to furnish wives for the eligible men. In 1 9 0 5 , for example, there were more than 2 0 "surplus" males for every 100 females in the age group 15-19, who could not find wives from the same age group, and would have difficulty in finding them from any younger groups of girls which had suffered similar high mortality (see Table 60). Apparently, as girls began to catch

PATTERNS OF MARRIAGE AND DIVORCE

up with the boys in the chances of surviving in earlier life, they lost some of this advantage relative to men in marrying; and, as the scarcity of marriageable women became less of a factor conditioning the marriage opportunities for men and women, the result was that the chances of marrying at each age (under, say, 30) improved for men and declined for women. TABLE 61 EXCESS OF MARRIED WOMEN IN TAIWAN, 1905-1935 Married Men

Married Women

Difference (2)-(1)

Males per IOOFemales

Taiwanese Japanese Other Total

594,827 12,302 3,487 610,616

598,556 8,246 220 607,022

3,729 -4,056 -3,267 -3,594

99.4 149.2 1,585.0 100.6

Taiwanese Japanese Other Total

641,746 37,377 7,427 686,550

648,764 32,946 2,270 683,980

7,018 -4,431 -5,157 -2,570

98.9 113.4 327.2 100.4

Taiwanese Japanese Other

913,655 56,562 15,342

924,963 52,663 8,132

11,308 -3,899 -7,210

98.8 107.4 188.7

Total

985,559

985,758

199

100.0

1905

1920

1935

Sources: Census of 1905, Detailed Tables, Table 5; Census of 1920, Descriptive Re­ port with Appendix, Appendix Table 26; Census of 1935, Statistical Tables, Table 7. The category of "Married" as published in 1920 has been recon­ stituted so as to make it the same as that shown in the other censuses.

All this would be strictly true only if the Taiwanese population had been "closed" against marriage with other groups as completely as it was against migration. There were barriers to intermarriage, but in fact they were less effective. To some extent women enjoyed their felicitous marriage prospects because their nuptial horizons were not entirely restricted to Taiwanese men. Some of them mar­ ried Mainland Chinese or even Japanese, which helped to enhance their scarcity. These mixed unions probably had a dubious legal status, for many of the male Japanese and Chinese nationals in Tai­ wan were little more than transients, and some who reported them-

PATTERNS OF MARRIAGE AND DIVORCE

selves as married may have done so because of a wife at home. It is unknown how many supplemental unions of this sort with Tai­ wanese women there were. They were not legally registered as mar­ riages where Japanese were involved, for prior to 1932 such mar­ riages were prohibited by law. The legal status of a union, however, did not make so much dif­ ference in the census figures on married persons, since the census was designed to record unregistered as well as "legal" marriages. In this it seems to have been successful, even for those that were not fully legitimized by law, and the census reports are therefore a fairly good source to detect whether or not mixed unions existed.2 A brief summary of this information is given in Table 61. With monog­ amous mating and without intermarriage, there should have been equal numbers of Taiwanese married men and married women at any one moment.3 Actually, every census showed an excess of women over men who were married: a proportion in the neighborhood of one per cent of Taiwanese women were "unmatched" in this sense. But although some of these had undoubtedly become wives of Japanese or Chinese nationals, from the census materials there is no way of determining how widespread this intermarriage was. In the first place, Taiwanese men of wealth were said to take concubines, after the fashion of old-style Chinese gentry, and thereby may have accounted for some of the extra married women.4 Also, from the Japanese (and Chinese) side, there is no way of knowing how many of these men were reported as married because of wives in Japan (or China), and how many because of wives among the Taiwanese. Finally, a small though unknown number of Taiwanese men had found wives from women in the other nationality groups in the island, and these have the effect of cancelling some of the "excess" of married Taiwanese women. These factors make the situation more complicated than it would ZThe first census (1905) was an exception. Liaisons with Japanese men and Tai­ wanese women were explicitly not recognized, though such unions involving Mainland Chinese were. In all other censuses, the "informally married"—i.e., those whose mar­ riages were not legally registered—were always shown separately, though sometimes in­ cluded as a subtotal of "married" and sometimes as a subtotal of "single." 3 Some Taiwanese men, by the time of the later censuses, were absent in Japan and elsewhere in the Empire, but they were too few to have any noticeable effect in these tabulations. 4 By the same token, the sex imbalance of married Taiwanese cannot be said to be a measure of concubinage, because of the unknown amount of intermarriage. Neither of these two questions, both important, can be clarified beyond this indeterminate state of information.

PATTERNS OF MARRIAGE AND DIVORCE

MARRIAGE RATES, TAIWANESE

FEMALE

"I·35

.30

1934-36

25

.25

MALE

UJ

2

.20

Φ

1919· 21

cc

cr o

U. O Z

>· 43"49> 193, 203ft Banks, 28, 29 Beal, E. G., Jr., 10, 76, 241, 267 Birth control, 262 Birth rates. See Fertility. Birth registration, 50, 240 Callis, H. G „ 264 Campbell, W., 134, 264 Camphor, 24, 31 Capital formation. See Investment. Census enumerations, 9-10, 57, 57n., 75-78, 173-174, 180-181, 214, 243 Census of 1940, citation, 57n. Ch'en Cheng-hsiang, viii, 5, 20, 35, 44, 265 Cheng Ch'eng-kung, 5 China, relations with, vii-viii, 4-5, 33, 259 Chinese population, ix, x, 148, 173. See also Taiwanese. Chinese terms, Glossary, 269 Cholera, 134-136 Cities composition, 117-132 fertility, 249, 258 growth, 1 1 4 - 1 1 7 health, 137 marriage, 233-235 migration, 109-114 modern influence, 1 1 5 , 1 1 7 - 1 1 8 , 125, 131132 Civil Affairs Handbook (U.S. Office of Naval Operations), 3 1 , 40, 50, 118, 122, 140, 267 Civil divisions, 102-104 Clan organization, 47, 48, i74n. Climate, 3, 135 Cohen, J., 32, 265 Colonial policies, ix, 7, i8ff., 49-52, 66-69, 256 Commerce, 23, 42, 46 Communications, 26, 29, 31, 47 Corporations, 24, 26-32, 41

Economic development government policies, 7, 2off. Japanese corporations, 19, 24, 26-32, 41 impact on agriculture, 18-19, 22 > 1 5 2 ® Economic expansion, 32ft, 56ft., 100-101 Economic organization. See Agriculture, Economic development, Family, Government administration, Investment. Education, 50, 68n., 8in., 138 Electricity, 29, 41 Employment patterns. See Occupations. Exports, 19-20, 24, 33 Factory employment, 41, 63-65 Family and livelihood, 79, 186-187, ig6ff., 205ff. descent, 18m., 184, 218, 229 influence of, 173ft. See also Marriage, Divorce, Households, Fertility. Farm. See Agriculture. Fertility, 238ft. differences, 248ft. inducements, 239 measurement, 241-246 natural increase, 135, 238 trend, 240 Fertilizer, in agriculture, 32, 37, 134 Government administration bureaucracy, 62, 65ft. Japanese, 24, 26ft., 49-52, 136-139 Manchu, 5-6, 23, 44-48, 135 Grajdanzev, A. J., viii, 20, 27, 28, 29, 265 Gross reproduction rates, 245ft. Growth of population, 4, 12-15, 106, 183 185, 255. See also Natural increase.

272

INDEX Hakka people, 4-5, 44 Health. See Disease. Households composition, i8off. definitions, 174, 176, 185 enterprises, 54 size, 1758:. Hsu Tze-wei, 50, 52, 121, 134, 138, 265 Hu Hsien-chin, 174, 265 Imports, 22, 37, 134 Increase of population. See Growth of population. Industries, 28, 32, 37, 4 1 , 59, 63 Infant mortality, 158-165, 244 infanticide, 135, i59n. Influenza, 146, 153 "Inheritance" of occupation, g4ff. Investment, 19, 27-28, 3 1 , 39-42, 62 Irrigation, 27, 35, 40-41, 135, 137 and double-cropping, 35, 37, 53 weapon of corporations, 30 Japan annexation of Taiwan, vii, 6-7 military campaign, 49, 136 World War II, 28-32, 4on. Japanese nationals in Taiwan, 10, 16, 43, 52, 65-70, 117-125, 135, 176-177, 180, 213, 238n. Kawada, Shiro, 44, 265 Kerr, G. H „ 68, 265 Koreans, 10, 16 Labor force and mortality, 153-155, 166 composition, 41, 56ft. definition, 74if. expansion, 61, 99-100 Land survey, 24, 25 Land tenures, 44, 47-48 Lang, O., 173, 210, 265 Leisure, 77n„ 197-198, 200-204 Levels of living, ig-20, 56 Levy, M. J., Jr., 173, 210, 266 L i f e expectancy, 153-154, 168 L i f e tables, 87, 141, 146-172 Local isolation, 18, 47, 134, 166

intermarriage, 2 1 3 remarriage, 217-218, 222ff. shortage of women, 212-213 Mendiola, N. B., 25, 26, 266 Migration, 17, 43, 68, 91, i02ff., 125 Mining, 63 Monopolies, 24, 31, 139 Mortality differences by place, 165-170 early, 14, 15 trend, 145ft. Naito, Hideo, 136, 266 Nasu, Shiroshi, 53, 266 Natural increase, 12, 13 and migration, 109 Notestein, F. W., 81, 245, 247, 266 Occupations changes, 56-72 cities, 130-132 definitions, 57, 58, 59, 75-78, 90 different groups, 65-72, 127-131, 183 movement, 62, 65-70, 84ft. Opium, 3 1 , 135, 139 OPNAV. See Civil Affairs Handbook. Order, maintenance of, 24-25, 44 Pao-chia system, 25, 50, 140, 159 Pickering, W. A., 45, 266 Plague, 134, 136 Police, 10, 50, 51, 137, 240 Population distribution, g (fig.), 104 Prefecture, 103 Public health early conditions, 24, 133-139 government policies, 24, 1 2 m . , 168-170, 171, 257

136s.,

Quo Sung-ken, 266 Registration system, viii, 9-12, go, 102, 139145. l 59 _ I 6°> 173. 240 Reubens, E. P., 22, 266 Rice production, 34, 35, 55-56 and sugar, 30

Malaria, 135, 137, 146, 166 Malaya, Chinese in, 248 Manpower. See Labor force, Factory employment. Markets, local scope, 28 Marriages, 16, 21 off. and reproduction, 23off.

Smith, T . E., 247, 248, 266 Social changes, 41-42, 56®., 102, 131-132, 133, 236-237, 260ft. Social organization. See Agriculture, Family, Cities, Marriages, Government administration. Statistical Summary, viii, 29, 30, 66 Statistics economic, 7, 32s. population. See Demographic data.

273

INDEX Sugar production, 24, 29-30, 32, 34, 48, 55 Taeuber, I. B., 10, 56, 76, 241, 245, 247, 267 Taiwan Development Company, 31 Taiwanese distinct from other Chinese, 11-12, 17, 103, 119, 177, 213 origins, x, 4-6, 11, 16, 43-44, 102, 232 separate group, 10-12, 17, 43, 123-124 Takekoshi, Yosaburo, 46, 47, 49, 134, 267 Taxes, 18, 24, 27, 50 Tea, 23, 35 Technical skills, 66ff. Topography, 3, 8 (fig.), 103 Trade, 20, 21, 29, 33, 46-47, 134, 135, 170

Tu Tsung-ming, 139, 267 Tuan Chi-hsien, viii, 265 United Nations, Population Division and Statistical Office, 74, 267 Urbanization. See Cities. Violence, 6, 15, 24, 47, 134 Vital statistics, See Demographic data, Winfield, G. F., 134, 268 Widows and widowers, 216-218 Women, position of, r.n, 79-80, 1 «a-ι K7, 182, 189, 193-196 Yanaihara, T., viii, 27, 50, 68, 268